Comparative mapping of disease resistance genes in Lathyrus spp. using model legume genetic information

Nuno Felipe Alves de Almeida

Dissertation presented to obtain the Ph.D degree in Biology Instituto de Tecnologia Química e Biológica António Xavier | Universidade Nova de Lisboa

Oeiras, July, 2015

PhD Supervisors

Maria Carlota Vaz Patto, PhD

Diego Rubiales, PhD

The work presented in this thesis was performed mainly at:

Under the supervision of Maria Carlota Vaz Patto, PhD

Under the supervision of Diego Rubiales, PhD

The student, Nuno Felipe Alves de Almeida received financial support from Fundação para a Ciência e Tecnologia and Fundo Social Europeu in the scope of Quadro Comunitário de Apoio through the PhD Fellowship SFRH/BD/44357/2008.

iii ISBN: 978-989-20-5841-2 Table of Contents

LIST OF FIGURES ...... XIII

LIST OF TABLES ...... XV

LIST OF ADDITIONAL FILES ...... XVII

ACKNOWLEDGMENTS / AGRADECIMENTOS ...... XIX

RESUMO ...... XXI

ABSTRACT ...... XXV

LIST OF ABBREVIATIONS ...... XXIX

CHAPTER 1 - GENERAL INTRODUCTION ...... 1

1.1. Introduction ...... 3 1.2. Lathyrus sativus and L. cicera: Origin and systematic ... 4 1.3. Lathyrus sativus varietal groups ...... 6 1.4. Genetic resources and utilization ...... 7 1.5. Major breeding achievements ...... 9 1.6. Specific goals in current breeding ...... 10 1.7. Biotic stresses...... 11 1.7.1. Rust ...... 12 1.7.2. Powdery mildew ...... 13 1.7.3. Ascochyta blight ...... 13 1.8. Breeding methods and specific techniques ...... 15 1.9. Integration of new biotechnologies in breeding programmes ...... 17

v Table of Contents ______

1.10. Objectives ...... 20 1.11. Acknowledgments ...... 21 1.12. References ...... 21

CHAPTER 2 - TRANSFERABILITY OF MOLECULAR MARKERS FROM MAJOR LEGUMES TO LATHYRUS SPP. FOR THEIR APPLICATION IN MAPPING AND DIVERSITY STUDIES ...... 31

2.1. Abstract ...... 33 2.2. Introduction ...... 34 2.3. Methods ...... 36 2.3.1. Plant Material and DNA isolation ...... 36 2.3.2. Microsatellite markers ...... 38 2.3.3. Intron-Targeted Amplified Polymorphic markers, Defence-Related Genes and Resistance Gene Analogues ...... 39 2.3.4. Sequencing ...... 40 2.3.5. Segregation and diversity studies ...... 40 2.3.6. Cleaved Amplified Polymorphic Sequence and derived Cleaved Amplified Polymorphic Sequence markers for mapping ...... 41 2.4. Results and discussion ...... 42 2.4.1. Transferability of microsatellites to Lathyrus spp...... 44 2.4.2. Transferability of Intron-Targeted Amplified Polymorphic markers, Defence-Related and Resistance Gene Analogues to Lathyrus spp...... 50 2.4.3. Usefulness of cross amplified markers on Lathyrus spp. linkage mapping ...... 52

vi Table of Contents ______

2.4.4. Usefulness of cross amplified markers on Lathyrus spp. diversity studies ...... 54 2.5. Conclusions ...... 59 2.6. Acknowledgements ...... 60 2.7. References ...... 60

CHAPTER 3 - ALLELIC DIVERSITY IN THE TRANSCRIPTOMES OF CONTRASTING RUST-INFECTED GENOTYPES OF LATHYRUS SATIVUS, A LASTING RESOURCE FOR SMART BREEDING ...... 69

3.1. Abstract ...... 71 3.2. Introduction ...... 72 3.3. Material and methods ...... 76 3.3.1. Plant and fungal material, inoculation and RNA isolation ...... 76 3.3.2. Sequencing and quantification ...... 77 3.3.3. SNP detection ...... 79 3.3.4. EST-SSR development and genotyping ...... 79 3.3.5. Quantitative RT-PCR assay ...... 80 3.3.6. Contig annotation and data analysis ...... 81 3.3.7. Availability of supporting data ...... 81 3.4. Results ...... 82 3.4.1. Contigs from the RNA-Seq transcriptomes of resistant and susceptible L. sativus genotypes ...... 82 3.4.2. RNA-Seq validation by quantitative RT-PCR assay ...... 83 3.4.3. Differential gene expression in resistant and susceptible L. sativus genotypes during infection ...... 84 3.4.4. Annotation ...... 87 3.4.5. Biotic stress related proteins ...... 92

vii Table of Contents ______

3.4.6. SNPs in resistance pathways ...... 96 3.4.7. EST-SRR development ...... 96 3.5. Discussion ...... 98 3.6. Conclusions ...... 105 3.7. Acknowledgments ...... 106 3.8. References ...... 106

CHAPTER 4 - LATHYRUS SATIVUS TRANSCRIPTOME RESISTANCE RESPONSE TO ASCOCHYTA LATHYRI INVESTIGATED BY DEEPSUPERSAGE ANALYSIS ...... 115

4.1. Abstract ...... 117 4.2. Introduction ...... 118 4.3. Material and methods ...... 120 4.3.1.1. Plant material and inoculation ...... 120 4.3.2. RNA extraction and deepSuperSAGE library construction ...... 120 4.3.3. Data analysis and annotation ...... 121 4.3.4. Quantitative RT-PCR assays ...... 122 4.3.5. UniTag assignment to functional categories ...... 123 4.4. Results ...... 125 4.4.1. DeepSuperSAGE library characterization ...... 125 4.4.2. DeepSuperSAGE validation by quantitative RT-PCR assay ...... 125 4.4.3. Annotation of differentially expressed genes in the resistant L. sativus genotype after A. lathyri infection ...... 126 4.5. Discussion ...... 128 4.5.1. Pathogen perception ...... 131

viii Table of Contents ______

4.5.2. Hormone signaling ...... 132 4.5.3. Cell wall fortification ...... 134 4.5.4. Antimicrobial activity ...... 137 4.5.5. Reactive oxygen species ...... 138 4.5.6. Detoxification ...... 138 4.6. Conclusions ...... 139 4.7. Acknowledgments ...... 140 4.8. References ...... 140

CHAPTER 5 - DIFFERENTIAL EXPRESSION AND ALLELIC DIVERSITY IN LATHYRUS CICERA / RUST INFECTION, A COMPREHENSIVE ANALYSIS ...... 147

5.1. Abstract ...... 149 5.2. Introduction ...... 149 5.3. Material and methods ...... 152 5.3.1.1. Plant material, inoculation and DNA and RNA isolation ...... 152 5.3.2. RNA sequencing and transcript quantification...... 153 5.3.3. Contig annotation and data analysis ...... 155 5.3.4. RNA-Seq validation by quantitative RT-PCR assay ...... 156 5.3.5. SNP detection ...... 157 5.3.6. Allele-specific expression analysis by dual labelled probe RT-qPCR assays ...... 157 5.3.7. EST-SSR development and genotyping ...... 158 5.4. Results ...... 158 5.4.1. RNA-Seq transcriptomes of analysed L. cicera genotypes...... 158 5.4.2. Differential gene expression in partially resistant and susceptible L. cicera genotypes to rust infection ...... 159

ix Table of Contents ______

5.4.3. RNA-Seq validation by quantitative RT-PCR assay ...... 161 5.4.4. Annotation of L. cicera contigs ...... 162 5.4.5. SNP identification in resistance pathways ...... 166 5.4.6. Allele-specific expression validation by dual probe RT-qPCR assays ...... 166 5.4.7. EST-SRR markers development ...... 168 5.5. Discussion ...... 168 5.6. Conclusions ...... 178 5.7. Acknowledgments ...... 178 5.8. References ...... 179

CHAPTER 6 - GENETIC BASIS OF PARTIAL RESISTANCE TO RUST AND POWDERY MILDEW IN LATHYRUS CICERA - QTL MAPPING AND SYNTENY STUDIES ...... 185

6.1. Abstract ...... 187 6.2. Introduction ...... 187 6.3. Material and methods ...... 191 6.3.1. Plant material ...... 191 6.3.2. Disease reaction evaluation ...... 191 6.3.2.1. Rust evaluation ...... 191 6.3.2.2. Powdery mildew evaluation ...... 192 6.3.3. DNA isolation and molecular markers screening ...... 193 6.3.4. Map construction ...... 194 6.3.5. Comparison with M. truncatula genome ...... 194 6.3.6. Phenotypic data analysis ...... 195 6.3.7. QTL mapping ...... 195 6.4. Results ...... 196

x Table of Contents ______

6.4.1. Disease evaluation under controlled conditions ...... 196 6.4.2. Linkage map construction ...... 198 6.4.3. Macrosynteny between L. cicera LGs and M. truncatula chromosomes ...... 200 6.4.4. QTL mapping for rust and powdery mildew resistance ...... 202 6.4.5. Discovery of potential candidate genes in the identified QTLs ...... 203 6.5. Discussion ...... 206 6.6. Acknowledgments ...... 210 6.7. References ...... 210

CHAPTER 7 - GENERAL DISCUSSION ...... 217

7.1. Project overview ...... 219 7.2. Molecular marker development ...... 221 7.3. Development of the first Lathyrus cicera linkage map ...... 223 7.4. Lathyrus spp. response to Uromyces pisi inoculation ...... 224 7.5. Powdery mildew resistance in Lathyrus cicera ...... 227 7.6. Lathyrus sativus response to Ascochyta lathyri inoculation ...... 228 7.7. Future research and practical applications ...... 229 7.8. References ...... 231

xi

List of Figures

Figure 2.1 Numbers and percentages of gSSRs and EST-SSRs from lentil, pea and faba bean amplified in Lathyrus spp...... 44

Figure 2.2 Numbers and percentages of ITAPs and pea DRs and RGAs amplified in Lathyrus spp...... 51

Figure 2.3 Neighbor-Joining tree based on the proportion-of- shared-alleles distance values among 20 Lathyrus individuals...... 58

Figure 3.1 Correlation between RNA-Seq and RT-qPCR...... 86

Figure 3.2 Venn diagram of the number of unique and shared contigs between the two genotypes and its expression...... 88

Figure 3.3 Number of contigs that could be BLASTed to different plant species...... 89

Figure 3.4 Percentage of contigs assigned in each main functional category...... 91

Figure 3.5 Percentage of contigs assigned in each functional category for each expression pattern group...... 93

Figure 3.6 Percentage of contigs containing SNPs between the resistant and susceptible genotypes in each Mercator mapping functional sub-category...... 97

Figure 4.1 Relative expression levels correlation between RNA- Seq and RT-qPCR...... 126

xiii List of Figures ______

Figure 4.2 Percentage of annotated up- and down-regulated L. sativus UniTags upon A. lathyri inoculation in several functional categories by MapMan...... 128

Figure 5.1 Venn diagram of the number of unique and shared contigs between the two genotypes and its expression...... 161

Figure 5.2 Number of contigs that could be BLASTed to different plant species...... 162

Figure 5.3 Percentage of contigs assigned in each main functional category...... 164

Figure 5.4 Percentage of contigs containing SNPs between the resistant and susceptible genotypes in each Mercator mapping functional sub-category...... 167

Figure 6.1 Frequency distributions of rust (experiment 2R) and powdery mildew (experiment 3PM) DS in the RIL families under controlled conditions...... 197

Figure 6.2 L. cicera first genetic linkage map based on a RIL population...... 201

Figure 6.3 Matrix plot of common gene-based SNP markers mapped in L. cicera and M. truncatula...... 202

Figure 6.4 MQM QTL mapping analysis for rust and powdery mildew resistance on LG II of the L. cicera linkage map developed...... 204

xiv List of Tables

Table 2.1 Lathyrus cicera and Lathyrus sativus accession references at the CRF-INIA germplasm bank...... 37

Table 2.2 Characteristics of developed CAPS and dCAPS markers...... 43

Table 2.3 Characteristics of pea microsatellite markers successfully amplified and sequenced on Lathyrus spp...... 46

Table 2.4 Lathyrus cicera BGE008277xBGE023542 RILs polymorphic markers characteristics and respective segregation analysis (χ2 test for 1:1 segregation)...... 53

Table 2.5 Lathyrus cicera and Lathyrus sativus cross amplifiable markers diversity analysis...... 56

Table 3.1 Log2 fold expression results for RNA-Seq and RT- qPCR experiments...... 85

Table 3.2 Classification of contigs according to their differential expression in the susceptible and resistant genotype upon infection with U. pisi...... 87

Table 4.1 Contig information and primer sequences for RT- qPCR...... 124

Table 4.2 Log2 fold expression results for deepSuperSAGE and RT-qPCR experiments...... 127

xv List of Tables ______

Table 4.3 List of detected genes by functional category, expression values and putative function as described in Mercator...... 129

Table 5.1 Classification of contigs according to their differential expression in the susceptible and resistant genotype upon infection with U. pisi...... 160

Table 5.2 Allele-specific expression analysis results...... 169

Table 6.1 Phenotypic values (mean ± standard deviation) of the RIL families and quantitative genetic parameters for rust and powdery mildew resistance...... 198

Table 6.2 Description of the obtained linkage groups ...... 200

Table 6.3 Quantitative trait loci for rust and powdery mildew resistance in a L. cicera RIL5 population (BGE023542xBGE008277)...... 203

Table 6.4 Potential candidate genes identified in QTLs...... 205

xvi List of Additional files

Chapter 3 Additional files can be found at http://www.biomedcentral.com/1471- 2229/14/376/additional. Please note that due to format changes, they appear under different names in the published version. Additional file 3.1 List of detected SNPs between grass pea genotypes BGE015746 and BGE024709. (published as Additional file 3) Additional file 3.2 List of detected polymorphic EST-SSRs between grass pea genotypes BGE015746 and BGE024709. (published as Additional file 4) Additional file 3.3 Contig information and primer sequences for RT- qPCR. (published as Additional file 5) Additional file 3.4 Lists of genes in expression pattern groups mapped to the reference assembly. RPKM: reads per kilobase per million (published as Additional file 1) Additional file 3.5 List of contigs, mapped to the reference assembly, present only in the inoculated condition with a higher bit-score in fungal databases than in plant databases. (published as Additional file 2)

Chapter 4 The Additional file can be found at http://www.frontiersin.org/journal/ 10.3389/fpls.2015.00178/abstract. Please note that due to format changes, it appears under a different name in the published version. Additional file 4.1 UniTags differentially expressed in Lathyrus sativus BGE015746 after inoculation with Ascochyta lathyri. TPM; tags per million.

xvii List of Additional files ______

Chapter 5 Additional files can be found at http://www2.itqb.unl.pt/~nalmeida/ external/thesis_chapter5/ Additional file 5.1 Log2 fold expression results for RNA-Seq and qRT-PCR experiments and primer sequences for qRT-PCR. Additional file 5.2 Transcript information for allele-specific RT-qPCR assays. Additional file 5.3 List of primer and probe sequences for the allele- specific expression analysis assay. Additional file 5.4 Lists of genes in expression pattern groups mapped to the reference assembly. RPKM: reads per kilobase per million. Additional file 5.5 List of contigs, mapped to the reference assembly, present only in the inoculated condition with a higher bit-score in fungal databases than in plant databases. Additional file 5.6 List of detected polymorphic SNPs between grass pea accessions BGE008277 and BGE023542 Additional file 5.7 List of detected polymorphic EST-SSRs between grass pea genotypes BGE008277 and BGE023542

xviii Acknowledgements Agradecimentos

I would like to express my gratitude to the people who directly or indirectly contributed in so many ways to this work. I would like to thank: - Maria Carlota Vaz Patto, my supervisor, for giving me the opportunity to start this PhD work and all the help and support throughout the last years. Your mentoring allowed me to improve at both professional and personal level. - Diego Rubiales, my co-supervisor, for all the guidance and support for this work. Your advices have been priceless. - Pedro Fevereiro, head of the BVC lab at ITQB, for the support and giving all the best conditions possible at his group. - Ana Maria Torres, for receiving me at her laboratory at IFAPA, and providing the proper resources and expertise for the cross-species amplification work. - Peter Winter, for receiving me at GenXPro and providing the logistics in my stay in Frankfurt. Your considerations were very important when I was trying to organize the avalanche of genomic data. - Zlatko Šatović and Manuela Veloso for the relevant advices as being part of my thesis committee. - Susana Leitão, at first for all the great help with the genotyping part of this work. Secondly, for having me as desk neighbour for the last 6 years and never losing the sanity (I hope). Now that you have your own PhD thesis to take care I wish you a pleasant time finishing it. Saudações Leoninas. - Mara Alves, for being the statistical master, helping in the statistical analysis, linkage map development and QTL mapping. Also,

xix Acknowledgements / Agradecimentos ______thank you for being the person in the lab who can understand my not- so-obvious (bad) jokes. Força aí com essa tese! - Inês Trindade and Rita Morgado, for the help in setting up, analysing and discussing the RT-qPCR assays. - All my former and present colleagues at the BCV lab (André Almeida, Catarina Bicho, Cátia, Clara, Diana, Jorge Paiva, Maria, Matilde, Olivia, Silvana, Sofia, Susana Araújo, Susana Neves, Susana Pera, Victor, Zé), for the science related discussions, but also for non- work related discussions. - All my colleagues at IAS-CSIC, that helped me in my short stays in Córdoba (particularly Alessio Cimino, Angel Aller, Angel Villegas, Elena Cabalero, Elena Prats, Eleonora, Estefania Carrillo, Javi Sanchez, Mariane, Nicolas, Pedro and Sara) with their expertise in the scientific topics and kindness to help the Portuguese guy to adapt to the city. - Ana Moral, por el precioso entrenamiento en la evaluación de resistencia y toda la preciosa ayuda dentro y fuera del laboratorio. - Thaïs, for the scientific and philosophical discussions. Also for sheltering me, making me feel at home. - All the people at GenXPro (particularly Björn, Conny, Günter, Nico and Ralf) for the contribution with the NGS part of my thesis and the help during my stay in Frankfurt. - Elsa and all the Kohl family, for the hospitality and friendship. - Aos meus pais, António e Liliana, por sempre me apoiarem. Este trabalho só é possível por me terem preparado para qualquer desafio. - À minha avó Emília, por ser sempre tão amável. - À minha filha Laura, por nos teres brindado com a tua alegria contagiante. A tua presença torna tudo mais brilhante. - À Dora, a minha amiga, companheira e pilar. Obrigado por me teres mantido focado e teres suportado a minha “ausência presente”.

xx Resumo

Entre as leguminosas de grão, Lathyrus sativus (chícharo) e L. cicera (chícharo-miúdo) detêm um grande potencial pela sua adaptabilidade a ambientes adversos, alto teor em proteína e resistência a doenças relevantes. As espécies do género Lathyrus são consideradas potenciais fontes de proteínas de alta qualidade e baixo custo. No entanto, devido à sua pouca utilização, esforços adicionais são necessários de forma a explorar o seu potencial e capitalizar os atuais avanços na biologia molecular para os programas de melhoramento em Lathyrus spp..

Nesta tese, a base genética dos mecanismos de defesa de duas espécies de Lathyrus, a três das mais importantes doenças foliares em leguminosas foram estudadas, nomeadamente, ferrugem (Uromyces pisi), oídio (Erysiphe pisi) e ascoquitose (Ascochyta lathyri). Dois genótipos de cada L. cicera e L. sativus com resposta contrastante à infeção com ferrugem e oídio, e um genótipo de L. sativus resistente à ascoquitose, foram utilizados para esta análise.

De forma a permitir a construção de um mapa de ligamento genético de L. cicera, tendo como base uma população de linhas puras recombinantes (RILs – recombinant inbred lines), foram desenvolvidos marcadores moleculares polimórficos. Duas abordagens diferentes foram utilizadas para o desenvolvimento de marcadores moleculares. Primeiro foram testados marcadores desenvolvidos para espécies próximas filogeneticamente de Lathyrus, como Medicago truncatula ou Pisum sativum. Apesar dos marcadores serem transferíveis entre espécies, obtiveram-se poucos marcadores polimórficos entre os genótipos parentais das RILs. De forma a superar este facto, foi

xxi Resumo ______efetuada uma segunda abordagem utilizando bibliotecas obtidas por sequenciação de ARN (RNA-Seq) de L. cicera e L. sativus (também desenvolvidas nesta tese), de forma a desenvolver marcadores (EST- SSR e SNPs) polimórficos específicos para estas espécies. Devido à incorporação no mapa de ligamento de L. cicera de vários marcadores homólogos em espécies modelo de leguminosas, foi possível efetuar estudos de sintenia. Este estudo indicou uma grande conservação macrosinténica entre L. cicera e M. truncatula, permitindo novas linhas de investigação associadas com o mapeamento comparativo de processos fisiológicos e mecanismos de defesa comuns. Beneficiando deste mapa de ligamento, também foi possível mapear locus de características quantitativas (QTL – quantitative trait locus) relacionadas com resistência a doenças. Foram detetados dois QTLs para a resistência parcial à ferrugem e um QTL para a resistência parcial ao oídio. A pequena percentagem de variação fenotípica total explicada pelos QTLs levou-nos a concluir que o controlo genético das resistências parciais ao oídio e à ferrugem é efetivamente poligénica.

Adicionalmente ao estudo genético, uma abordagem usando transcriptómica (RNA-Seq) foi efetuada para ambas as espécies de forma a elucidar quais as respostas defensivas da planta à infeção por ferrugem. Os perfis de transcrição de L. sativus revelaram diferenças consideráveis na regulação das vias de sinalização hormonal mais importantes entre o genótipo resistente e o suscetível. Além disso, vários genes relacionados com patogenicidade foram sobre- expressos no genótipo resistente e sub-expressos no genótipo suscetível.

Os resultados de transcriptómica de L. cicera sugerem uma regulação diferencial de genes envolvidos na sinalização, metabolismo da parece celular e na síntese de metabolitos

xxii Resumo ______secundários como base genética da resistência parcial à ferrugem. Particularmente, um homólogo do gene PsMLO1 encontrava-se diferencialmente expresso após inoculação com ferrugem. Este gene já havia sido descrito como estando envolvido na resistência ao oídio em P. sativum e o seu papel na resistência de L. cicera à ferrugem deve ser melhor estudada. Os genes identificados como diferencialmente expressos são genes candidatos adequados a futuros estudos funcionais, de forma a esclarecer os mecanismos das interações planta-patógeno. Adicionalmente, as duas espécies de Lathyrus possuíam milhares de contigs polimórficos entre seus genótipos, com SNPs distribuídos de forma desigual entre as diferentes categorias funcionais. As categorias mais mutadas foram degradação de proteínas e proteínas cinase recetoras envolvidas na sinalização, o que ilustra a adaptação evolutiva destas espécies no braço de ferro entre hospedeiro/patógeno.

Uma abordagem transcriptómica diferente, deepSuperSAGE, foi também utilizada para elucidar as vias diferencialmente reguladas e identificar candidatos a genes de resistência na interação ascochyta- L. sativus. Os resultados indicam que varias classes de genes, atuando em diferentes fases da interação planta-patógeno, estão envolvidos na resposta de L. cicera à infeção por A. lathyri. Por exemplo, foi observada uma clara sobre-expressão de genes relacionados com defesa envolvidos na via biosintética do etileno. Houve também evidências de alterações no metabolismo da parede celular, indicada pela sobre-expressão de genes envolvidos na biossíntese de celulose e lignina.

Juntando todos os dados de transcriptómica e mapeamento de QTLs, estes resultados fornecem uma visão global dos perfis de expressão génica dos genótipos de Lathyrus spp. inoculados com

xxiii Resumo ______ferrugem, oídio e ascoquitose, fornecendo recursos muito importantes para abordagens futuras usando o melhoramento de precisão nestas valiosas leguminosas, até agora pouco estudadas.

xxiv Abstract

Lathyrus cicera L. (chickling pea) and L. sativus L. (grass pea) have great potential among grain legumes due to their adaptability to inauspicious environments, high protein content and resistance to serious diseases. Lathyrus spp. are considered potential sources of high quality and cheap protein. Nevertheless, due to its past underuse, further activities are required to exploit this potential and to capitalise on the present molecular biology advances on Lathyrus spp. breeding programmes.

In this thesis the genetic basis of the defence mechanisms, of two Lathyrus spp. to three of the most important foliar diseases in legumes, rust (Uromyces pisi), powdery mildew (Erysiphe pisi) and ascochyta blight (Ascochyta lathyri) were studied. Contrasting genotypes of both L. sativus and L. cicera in what concerns infection reaction to rust and powdery mildew, and a resistant L. sativus genotype against ascochyta blight were used in this analysis.

Polymorphic molecular markers that enabled the construction of a L. cicera linkage map base on recombinant inbred lines population were developed. Two different approaches were used in this molecular markers development. First we tested markers developed for Lathyrus close related species, such as Medicago truncatula and Pisum sativum that despite a good transferability, yielded a low amount of polymorphic markers between the RILs parental genotypes. To overcome that, and as a second approach we used RNA-Seq libraries of L. cicera and L. sativus (also developed in this thesis) to develop specific polymorphic EST-SSRs and SNPs. Due to the incorporation of several homologous

xxv Abstract ______markers to model legume species in the developed L. cicera linkage map, it was possible to perform synteny studies. This indicated a high macrosyntenic conservation between L. cicera and M. truncatula, opening research opportunities associated with comparative mapping of shared physiological process and defence mechanisms. Profiting from this linkage map, we also evaluated the L. cicera RILs for rust and powdery mildew resistance response in order to detect and map QTLs underlying disease resistance. One QTL for partial resistance to powdery mildew and two QTLs for partial resistance to rust were detected. The small percentage of total phenotypic variation explained by the detected QTLs led us to conclude that the genetic control of the partial resistances to rust and powdery mildew was indeed polygenic.

In addition to the genetic study, a transcriptomics approach (RNA-Seq) was used for both species to elucidate the defence responses to rust infection. L. sativus, transcription profiles revealed considerable differences in regulation of major phytohormone signalling pathways between resistant and susceptible genotypes. Also, several pathogenesis-related genes were up-regulated in the resistant and exclusively down regulated in the susceptible genotype. L. cicera transcriptomic results suggested different regulation of genes involved in signalling, cell wall metabolism and in the synthesis of secondary metabolites as the genetic basis of partial resistance to rust. In particular a PsMLO1 homolog was found differentially expressed upon inoculation with rust. This gene was already described as involved in powdery mildew resistance in pea, and its role in L. cicera rust resistance should be further investigated. The differentially expressed genes identified are suitable candidate genes for future functional studies to shed light on the molecular mechanisms of plant- pathogen interactions. In addition, the two Lathyrus spp. contained

xxvi Abstract ______thousands of polymorphic contigs between each species genotype, with SNPs unevenly distributed between different functional categories. Protein degradation and signalling receptor kinases were the most mutated categories, illustrating evolutionary adaptation of L. sativus to the host/pathogens arms race.

A different transcriptomic approach, deepSuperSAGE, was also employed to elucidate the pathways differentially regulated and identify resistance candidate genes during ascochyta-L. sativus interaction. The results indicated that several gene classes acting in different phases of the plant/pathogen interaction are involved in the L. sativus response to A. lathyri infection. As example a clear up- regulation of defence-related genes related with the ethylene pathway was observed. There was also evidence of alterations in cell wall metabolism indicated by overexpression of cellulose synthase and lignin biosynthesis genes.

Taking all the transcriptomics data and QTL mapping together, our results provide a broad overview of gene expression profiles of Lathyrus spp. genotypes inoculated with rust and Ascochyta, providing a highly valuable resource for future smart breeding approaches in this hitherto under-researched, valuable legume crop.

xxvii

List of most used abbreviations

ABA abscisic acid

AFLP Amplified Fragment Length Polymorphisms

CAPS Cleaved Amplified Polymorphic Sequence d.a.i. days after inoculation

DAMP damage-associated molecular pattern dCAPS derived Cleaved Amplified Polymorphic Sequence ddPCR droplet digital PCR

DE differentially expressed

DR disease resistance

DS disease severity eQTL expression quantitative trait locus

ER endoplasmic reticulum

EST expression sequence tags

EST-SSR expressed sequence tag - simple sequence repeat

ET ethylene

ETI effector-triggered immunity gSSR genomic simple sequence repeat

GST glutathione S-transferase h.a.i. hours after inoculation

HR hypersensitive response

IT infection type

ITAP intron-targeted amplified polymorphism

ITS internal transcribed spacer

xxix List of most used abbreviations ______

JA jasmonate

LG linkage group

LOD logarithm (base 10) of odds

LRR Leucine-rich repeat

MAP mitogen-activated protein

MAS marker assisted selection

MLO mildew resistance locus O

MQM multiple-QTL mapping

NBS-LRR nucleotide-binding site leucine-rich repeat

NGS next-generation sequencing

ODAP ß-N-ozalyl-L-a,ß-diaminopropanoic acid

PAMP pathogen associated molecular patterns

PCR Polymerase Chain Reaction

PIC Polymorphic Information Content

PR pathogenesis related

PRR pattern recognition receptor

PTI pathogen-associated molecular pattern triggered immunity

QTL quantitative trait locus

R gene resistance gene

RAPD Random Amplified Polymorphic DNA

RGA resistance gene analogue

RIL Recombinant Inbred Line

RNA-Seq RNA-Sequencing

ROS reactive oxygen species

RT reverse-transcription

xxx List of most used abbreviations ______

RT-qPCR Real-Time Quantitative Reverse Transcription PCR

SA salicylic acid

SNP single nucleotide polymorphism

SSR simple sequence repeats, or microsatellite

VLCFA very long chain fatty acid

xxxi

CHAPTER 1

General introduction

This chapter is based on: Almeida, N.F., Rubiales, D., and Vaz Patto, M.C. (2015). Grass pea. In: De Ron, A.M. (Ed.) Grain Legumes, Series Handbook of Plant Breeding. (New York, U.S.A.: Springer Science+Business Media). (in press) (see Acknowledgements section for authors contributions)

1

General introduction ______

1.1. Introduction

Lathyrus sativus L. (grass pea) and L. cicera L. (chickling pea) are multipurpose robust cool season legume crops. They can grow in both drought- and flooding-prone environments and poor soils due to its hardy and penetrating root systems (Campbell, 1997; Vaz Patto et al., 2006b). They have high nutritional value (protein content raging 25–30%), being grass pea important both for human food and animal feed, while chickling pea is more usually used as animal feed and forage. In what concerns human consumption grass pea can be consumed uncooked as a green snack, cooked in a stew, milled into flour or by roasting the seed (Peña-Chocarro and Peña, 1999). In addition to its uses as food and feed, symbiosis with rhizobia allows an efficient nitrogen fixation in the soil, lowering the inputs needed in crop rotation and making them suitable to be used as green manure in sustainable farming systems (Hanbury et al., 2000). As an example of its versatility, L. sativus is easily introduced in intercropping systems, rotations or used along with paddy rice in relay cropping systems (Campbell et al., 1994; Abd El Moneim et al., 2001; Hillocks and Maruthi, 2012).

There is great potential for the expansion in the utilization of Lathyrus spp. in dry areas or zones which are becoming more drought- prone, with increased salinity or increased tendency to suffer from biotic stresses. However, those species, and in particular grass pea, are unpopular with governments and donors because they contain small amounts of a toxin, β-N-ozalyl-L-α,β-diaminopropanoic acid (ODAP). Although this toxin can cause a neuronal disorder, known as ‘lathyrism’, the condition develops in humans with a 6% chance only when grass pea is consumed in large quantities, unaccompanied by other foodstuffs in an unbalanced diet and during a long period of time

3 Chapter 1 ______

(Lambein et al., 2009). Also, seeds can be partly detoxified by the various processing methods (Kuo et al., 2000; Kumar et al., 2011).

Even though, these robust crops are rightly considered as a model crop for sustainable agriculture and despite the lathyrism stigma, the development of new breeding technologies and the growing interest in its use in Mediterranean type environments, all over the world will provide a bright future to this crop (Vaz Patto et al., 2006b; Vaz Patto and Rubiales, 2014a).

1.2. Lathyrus sativus and L. cicera: Origin and systematic

The Lathyrus genus is located within the Fabaceae family (syn. Leguminosae), subfamily Faboideae (syn. Papilionoideae), tribe Fabeae (syn. Vicieae), along with genera Pisum, Vicia, Lens and Vavilovia (Wojciechowski et al., 2004; Kenicer et al., 2005; Smýkal et al., 2011; Schaefer et al., 2012).

The natural distribution of grass pea has been completely obscured by its human cultivation. Its use for food, feed and forage difficult the distinction between wild and domesticated populations, toughen the task to precisely locate its centre of origin (Kumar et al., 2013). The most probable grass pea centre of origin is believed to have been the Eastern Mediterranean or Fertile Crescent, around 6.000 B.C.E. This has been supported by archeobotanical and recent phylogenetic reports (Kislev, 1989; Schaefer et al., 2012), refuting the hypothesis by Smartt (1984) that the centre of origin was located in south-west or central Asia. Domestication of grass pea seems to have occurred alongside with other pulses, being normally found with early domesticates of pea (Pisum sativum L.), lentils (Lens culinaris Medik.) and bitter vetch (Vicia ervilia (L.) Willd.) (Erskine et al., 1994).

4 General introduction ______

Hopf (1986) hypothesized that L. sativus is a derivative from L. cicera, its genetically nearest wild species. In addition, in what concerns domestication in Southern Europe (France and Iberian peninsula), evidences of cultivation of L. cicera were found, dating from 4.000 or 3.000 B.C.E., suggesting that expansion of L. sativus farming may have led also to the domestication of the local L. cicera (Campbell, 1997).

Within the economically important legume crops and model species, P. sativum is reported as the closest related to grass pea and chickling pea, followed by lentil, faba bean (Vicia faba L.), barrel medic (Medicago truncatula Gaertn.), chickpea (Cicer arietinum L.) and Lotus corniculatus L. (Asmussen and Liston, 1998; Wojciechowski et al., 2004; Ellison et al., 2006).

The infrageneric classification of Lathyrus genus has been revised several times, being the one reported by Kupicha (1983) the most largely accepted. In this treatment, the genus is organized in 13 clades (Orobus, Lathyrostylis, Lathyrus, Orobon, Pratensis, Aphaca, Clymenum, Orobastrum, Viciopsis, Linearicarpus, Nissolia, Neurolobus, and Notolathyrus). This morphological based classification has been recently supported by molecular phylogenetic studies using sequence data from the internal transcribed spacer (ITS) region and from cpDNA (Kenicer et al., 2005; Kenicer et al., 2009). Schaefer et al. (2012), using nuclear and chloroplast phylogenetic data, further suggested that the genus Lathyrus is not monophyletic, and recommended that a more natural classification would be to transfer Pisum and Vavilovia to a then monophyletic Lathyrus genus.

5 Chapter 1 ______

1.3. Lathyrus sativus varietal groups

Great morphological variation is reported in grass pea, especially in vegetative characters such as leaf length, while, for instances, its floral characters are much less variable, showing a clear grouping in flower colour (Jackson and Yunus, 1984), as well as in its seed and yield traits (Hanbury et al., 1999). Several studies divided grass pea accessions broadly into two groups; those from the Indian sub-continent and those from the Mediterranean region. Jackson and Yunus (1984) reported that all blue-flowered accessions came from south-west and south Asia, while the white and mixed coloured accessions had a more western distribution, from the Canary Isles to the western republics of the Soviet Union. These authors also pointed out that white flowered accessions only had white seeds with no secondary markings on the seed coat. In accordance with this, Hanbury et al. (1999), reported that Mediterranean accessions were characterized by larger and whiter seeds, selected for human consumption, with higher yield potential than the Indian accessions. Grass pea small-seeded accessions are considered more primitive types and normally associated with hardened seeds like what happens in other Old World grain legumes such as pea, chickpea or lentil (Chowdhury and Slinkard, 2000).

A particular case is the germplasm selected for forage, in the Mediterranean region, with landraces with broad leaves and pods, but low seed yield (Chowdhury and Slinkard, 2000; Kumar et al., 2013).

6 General introduction ______

1.4. Genetic resources and utilization

Conservation of Lathyrus genetic resources has recently attracted more attention because of the potential role of these species under the climate change scenario (Kumar et al., 2013).

Grass pea is mentioned in two conservation programs for major food legumes. One is the International Treaty on Plant Genetic Resources for Food and Agriculture (ITPGRFA) (FAO, 2009), which aims at guaranteeing food security through conservation of biodiversity, fair exchange and sustainable use of plant genetic resources. This is being accomplished by establishing a global system to provide farmers, plant breeders and scientists’ access to plant genetic materials, ensuring that recipients share benefits with the countries where they have been originated and by recognizing the contribution of farmers to the diversity of crops used as food.

The other, a more specific program developed by the Global Crop Diversity Trust (CGDT) in collaboration with ICARDA, aims for a long-term conservation strategy of L. sativus, L. cicera and L. ochrus (GCDT, 2009). This program is detailing the current status of national collections and identifying gaps in collections of these three species from areas of diversity. Their strategy recommends that documentation on collections should be upgraded and that more work should be carried out on characterizing and evaluating collections for key traits, making this data widely available (Gurung and Pang, 2011).

Several ex situ and a few in situ conservation examples exist for Lathyrus germplasm. The largest Lathyrus ex situ collections are maintained at the Conservatoire Botanique National des Pyrénées et de Midi-Pyrénées in France (4.477 accessions) (previously at Pau University), by the International Center for Agricultural Research in Dry

7 Chapter 1 ______

Areas (ICARDA) comprising 3.239 accessions, and by the National Bureau of Plant Genetic Resources (NBPGR) in India (2.619 accessions). Smaller, but still relevant collections are maintained by other banks such as the Germplasm Resource Information Network (GRIN) from the United States Department of Agriculture (USDA) in the United States of America, the Leibniz Institute of Plant Genetics and Crop Plant Research (IPK) in Germany and the Centro de Recursos Fitogenéticos (CRF) from the Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA) in Spain. Backups from 2.134 grass pea accessions, from 44 countries, and 176 chickling pea accessions, from 20 countries, are deposited at the Svalbard Global Seed Vault (http://www.nordgen.org/sgsv/; accessed January 2015). In what concerns in situ conservation, five genetic reserves for Lathyrus diversity conservation have been proposed in Syria and Turkey (Heywood et al., 2007). These authors also stressed the importance of increasing public awareness for the significance of crop wild relatives in agricultural development and the need for their simultaneous conservation.

This conserved germplasm represent a valuable reservoir of diversity, providing access to sources of a wide range of interesting agromorphological traits such as earliness, plant architectural traits, disease and pest tolerance, as well as low ODAP content. Characterization of this diversity through phenotyping and genotyping studies will unveil novel alleles that can be used to improve this crop. Diversity characterization in Lathyrus germplasm have focused for example on ODAP content (Fikre et al., 2008; Kumar et al., 2011; Grela et al., 2012), phenology and yield (Mera, 2010; Grela et al., 2012), parasitic weed resistance (Fernández-Aparicio et al., 2012), disease resistance (Gurung et al., 2002; Vaz Patto et al., 2006a; Vaz Patto and

8 General introduction ______

Rubiales, 2009) or quality traits (Granati et al., 2003). Some of these characterization studies have represented the first steps of several existing selection programs.

1.5. Major breeding achievements

Conventional grass pea breeding programs have been established in several countries, including Australia (Hanbury et al., 1995), Bangladesh (Malek, 1998), Canada (Campbell and Briggs, 1987), China (Yang and Zhang, 2005), Chile (Mera et al., 2003), Ethiopia (Tadesse and Bekele, 2003), India (Lal et al., 1986; Pandey et al., 1996), Nepal (Yadav, 1996) Syria (Abd El-Moneim et al., 2000) and Portugal (Carita, 2012). Some of these breeding programs are still active, but most are small in comparison to other legume crops (Vaz Patto et al., 2011).

Due to the occurrence of lathyrism in humans, major breeding programs essentially aimed for low ODAP content, besides productivity and adaptability. This has resulted at present in several L. sativus or L. cicera breeding lines or released varieties with reduced ODAP content (from 0.5 to 1.5 %, down to 0.01 % or less) (Kumar et al., 2011). For instance, low ODAP grass pea cultivars have been released in several countries, such as “Wasie” in Ethiopia, “Ali-Bar” in Kazakhstan and “Gurbuz 1” in Turkey (ICARDA, 2006; 2007). Similarly, low ODAP and high yielding grass pea cultivars have been released in India such as “Pusa 24”, “Prateek”, “Ratan” and “Mahateora” (ICAR, 2009). In Bangladesh, examples are the low ODAP and high-yielding grass pea varieties “BARI Khesari 1”, “BARI Khesari 2” and “BARI Khesari 3” (Malek 1998), or the “BINA Khesari 1” (Kumar et al., 2011). In Canada, high yield and low ODAP (0.03%) grass pea variety “LS8246” was

9 Chapter 1 ______released for feed and fodder (Campbell and Briggs, 1987), in addition to a high N fixation grass pea variety, “AC Greenfix”, released specially as green manure (Krause and Krause, 2003). In Chile, “Luanco-INIA”, a large-seeded, high yielding grass pea variety was released, used locally as feed and for export, especially for some European markets where larger seed size is desirable for human consumption (Mera et al., 2003). Finally, in Australia, the grass pea variety “Ceora” was bred to be used as forage, hay or as a green manure crop (Siddique et al., 2006). Also in Australia a chickling pea cultivar, “Chalus”, was selected for high yields and low ODAP levels (Hanbury and Siddique, 2000). In Portugal two chickling pea varieties are registered in the “Catálogo Nacional de Variedades” (http:// www.dgv.min-agricultura.pt/xeov21/attachfileu.jsp?look_parentBoui= 4259527&att_display=n&att_download=y, accessed September 2014), named “Grão da Comenda” and “Grão da Gramicha”, both to be used as forage.

1.6. Specific goals in current breeding

Low ODAP content is still one important goal of many of the current Lathyrus spp. breeding programs. Nevertheless other important agronomic traits have always been associated to this in breeding programs, such as yield and adaptation.

Increased yield is a selection criterion for most crop improvement programmes. However, some of the yield components that affect yield, such as double podding or increased seeds per pod, have in Lathyrus spp. breeding received insufficient attention. Also the biomass yield of L. sativus has started to receive more attention during the past few years (Campbell, 1997; Abd El Moneim et al., 2001; Vaz Patto et al., 2006b). This is a very important area due to the large

10 General introduction ______potential of this crop for forage and straw in the North African and South Asian regions (Campbell, 1997). Additionally, undesirable traits such as prostrate plant habit, indeterminate growth, late maturity and pod shattering (Rybinski, 2003) are being also handled by several breeding programs.

The concentrated effort on reducing ODAP content resulted in many other areas of evaluation and crop improvement, such as resistance to biotic and abiotic stresses, being neglected. However, with the release of low ODAP lines, the development of varieties, for instance, with increased resistance to prevalent pests and diseases has gained new strength. This crop is usually grown by poor farmers and under poor management, where it is difficult to adopt chemical control for diseases and pests. Therefore, the development of varieties having resistance to prevalent biotic stresses is essential and more efforts are required in this area of improvement of these very hardy crops (Vaz Patto et al., 2006b).

1.7. Biotic stresses

Grass pea and chickling pea as any other plant species are subjected to diseases caused by a vast array of pathogens, including fungi, viruses, bacteria, parasitic plants and insects. Previous studies identified resistance in grass pea and chickling pea germplasm for rust (Vaz Patto et al., 2009; Vaz Patto and Rubiales, 2009), powdery mildew (Vaz Patto et al., 2006a; 2007), ascochyta blight (Gurung et al., 2002), bacterial blight (Martín-Sanz et al., 2012) and crenate broomrape (Sillero et al., 2005; Fernández-Aparicio et al., 2009; Fernández-Aparicio and Rubiales, 2010; Fernández-Aparicio et al., 2012). However, only the genetic basis of aschocyta blight resistance

11 Chapter 1 ______in L. sativus was analysed through a quantitative trait locus (QTL) mapping and expression analyses of a few candidate genes (Skiba et al., 2004a; b; 2005), and detailed molecular information is missing for the majority of the identified resistances, hampering their introduction in breeding programs.

This thesis will focus on three of the most important fungal diseases for legume crops in which resistance was identified in Lathyrus spp.

1.7.1 Rust

Rusts are among the most important diseases of legumes (Sillero et al., 2006) and Lathyrus spp. are not exceptions (Duke, 1981; Campbell, 1997; Vaz Patto et al., 2006b). Rusts are caused by biotrophic fungi that keep infected host cells alive for their development, depending on the hosts to reproduce and complete their life cycles. Although some rusts can be cultured on very complex synthetic media, they have no known saprotrophic existence in nature (Staples, 2000). Rusts form elaborate intracellular feeding structures called haustoria, which maintain an intimate contact between fungal and plant cells over a prolonged period of time (O’Connell and Panstruga, 2006).

Rust in Lathyrus spp. is caused by Uromyces pisi (Pers.) Wint. and U. viciae-fabae (Pers.) J. Schröt. (Barilli et al., 2011; Barilli et al., 2012). In particular, U. pisi infects a broad range of other legumes, such as Vicia faba, Lens culinaris, Vicia ervilia and Cicer arietinum (Barilli et al., 2012; Rubiales et al., 2013).

The resistance observed in L. cicera and L. sativus against rust infection is due to a restriction of haustoria formation with high

12 General introduction ______percentage of early aborted colonies, reduction of number of haustoria per colony and reduction of intercellular growth of infection hyphae (Vaz Patto et al., 2009; Vaz Patto and Rubiales, 2009; Vaz Patto and Rubiales, 2014b).

1.7.2 Powdery mildew

Powdery mildews are probably the most common, conspicuous and widespread plant diseases. As biotrophs they seldom kill their hosts, but utilize their nutrients, reduce photosynthesis, increase respiration and transpiration, impair growth and reduce yields up to 40% (Agrios, 2005). Erysiphe pisi DC. is a biotrophic ascomycete fungus, characterized by its grey to white colonies formed on leaves, stem and pods of infected plants (Vaz Patto et al., 2006a). It is commonly known as pea powdery mildew but it can also infect Medicago, Vicia, Lupinus, Lens and Lathyrus (Sillero et al., 2006). Pea powdery mildew is a serious disease of worldwide distribution, being particularly important in climates with warm, dry days and cool nights (Smith et al., 1996).

Lathyrus sativus and L. cicera accessions with reduced disease severity despite of a high infection type after E. pisi infection, have also been identified (Vaz Patto et al., 2006a; 2007; Vaz Patto and Rubiales, 2014b), fitting the definition of Partial Resistance according to Parlevliet (1979).

1.7.3 Ascochyta blight

Ascochyta blights are among the most important groups of plant diseases worldwide (Rubiales and Fondevilla, 2012). Ascochyta blights

13 Chapter 1 ______are incited by different pathogens in the various legumes. As examples, Ascochyta rabiei (teleomorph Didymella rabiei) in chickpea; A. fabae (teleomorph D. fabae) in faba bean and A. lentis (teleomorph D. lentis) in lentil (Tivoli et al., 2006). Ascochyta blight of pea (Pisum sativum) is caused by a complex of fungi formed by Ascochyta pisi Lib., Didymella pinodes (Berk. & Blox.) Petrak. and Phoma medicaginis var. pinodella (L.K. Jones) Morgan-Jones & K.B. Burch. (Carrillo et al., 2013). Of these, D. pinodes (syn. Mycosphaerella pinodes) is the most frequent and damaging (Tivoli and Banniza, 2007). D. pinodes is a necrotrophic pathogen, being characterized by dark concentric lesions containing black picnidia on leaves, pods and stems (Peever, 2007). Lathyrus spp. are known to be resistant to D. pinodes, the causal agent of pea ascochyta blight. Gurung et al. (2002) showed that L. sativus, L. ochrus and L. clymenum accessions were significantly more resistant to D. pinodes stem infection than field pea cultivars. A detailed analysis of quantitative resistance of L. sativus to ascochyta blight, caused by D. pinodes, suggested that resistance in L. sativus may be controlled by two independently segregating genes, operating in a complementary epistatic manner (Skiba et al., 2004b). In another study, Skiba et al. (Skiba et al., 2004a) developed a grass pea linkage map and used it to locate two QTL, explaining 12% and 9% of the observed variation in resistance to D. pinodes. Nevertheless, no candidate genes were identified at that time for these resistance QTLs, hampering their use in precision breeding. In an attempt to identify defence-related candidate genes involved in D. pinodes resistance in L. sativus, the expression of 29 potentially defence-related expression sequence tags (ESTs) was compared between L. sativus resistant and susceptible lines (Skiba et al., 2005). These ESTs were selected from a previously developed cDNA library of L. sativus stem and leaf tissue challenged with D. pinodes. From these, sixteen ESTs were

14 General introduction ______considered eventually important for conferring stem resistance to ascochyta blight in L. sativus. In addition, the marker developed from one of them, EST LS0574 (Cf-9 resistance gene cluster), was significantly linked to one of the previously identified resistance QTLs. However this study was necessarily limited to the small number of initially selected EST sequences.

1.8. Breeding methods and specific techniques

Collection and evaluation of germplasm, local or introduced, is the cornerstone in any breeding program. Subsequent hybridization and selection of the resulting progeny using different strategies, will allow incorporating interesting traits into more adapted background. This may include backcrossing, recurrent selection, single seed descent and pedigree/bulk breeding methods. All of these methods can be applied on Lathyrus spp. improvement.

Grass pea and chickling pea are predominantly self-pollinated crops, although outcrossing up to 30% has been reported (Rahman et al., 1995; Chowdhury and Slinkard, 1997; Ben Brahim et al., 2001). Large size of flower, bright colour of petals, flower density, and nectar production are reported to influence the outcrossing in Lathyrus species (Kiyoshi et al., 1985). Entomophilic pollination in grass pea is due especially to bees and bumblebees (Kumar et al., 2011). Due to this observed outcrossing level, in most Lathyrus spp. breeding programmes, crosses are done under controlled conditions, in greenhouse or under insect proof coverings (Vaz Patto et al., 2011).

Conventional grass pea breeding focused essentially in hybridization of selected accessions, with the screening and evaluation of the resulting progeny. In the particular case of breeding to reduce

15 Chapter 1 ______

ODAP contend, low ODAP accessions are crossed with high yield material with good agronomic potential (Campbell, 1997).

Intergeneric hybridization, although difficult, is possible between grass pea and L. amphicarpos or L. cicera (Yunus and Jackson, 1991). Crosses have been also made with other species such as L. chrysanthus, L. gorgoni, L. marmoratus and L. pseudocicera (Heywood et al., 2007), but only ovules were produced.

Also with the objective of reducing ODAP content, grass pea has been subjected to induced mutagenesis by physical and/or chemical mutagens. Other traits have been affected by mutagenesis such as plant habit, maturity, branching, stem shape, leaf size, stipule shape, flower colour and structure, pod size, seed size and colour and NaCl tolerance (Nerkar, 1972; 1976; Rybinski, 2003; Biswas, 2007; Talukdar, 2009; 2011). In vitro culture was also employed, inducing somaclonal variation (Roy et al., 1993; Ochatt et al., 2002a; Zambre et al., 2002). Induced mutagenesis and somaclonal variation created new diversity, allowing the selection of lines with interesting agronomical traits, such as yield, plant architecture and low ODAP content,

Ochatt et al. (2002b) developed an in vitro system coupled with in vivo stages in order to shorten grass pea regeneration cycles, obtaining up to almost 4 cycles per year. However this approach is only applicable when few seeds/plant are intended, as in single-seed descendant breeding schemes.

The advent of various molecular marker techniques and the ability to transfer genes across different organisms, using transgene technology, has begun to have an impact on plant genome research and breeding. These techniques offer new approaches for improving important agronomic traits in Lathyrus species and breaking down

16 General introduction ______transfer barriers to related legume species (Vaz Patto et al., 2006b). This would allow exploring the variability existing in other Lathyrus gene pools and hopefully transfer the interesting grass pea and chickling pea traits to related legume species.

Genetic transformation of grass pea was attempted with only one successful report obtaining stable transformed plants (Barik et al., 2005). Given that regeneration protocols are often genotype specific, it may be necessary either to develop more generally applicable protocols or to adapt the protocol after transformation (Ochatt et al., 2013).

1.9. Integration of new biotechnologies in breeding programmes

Comparing to other grain legumes such as pea, faba bean or chickpea, genomic recourses for grass pea are still scarce. Prior to the inclusion of this thesis results, the NCBI database had available the information of 178 EST sequences from a cDNA library of one L. sativus accession inoculated with Mycosphaerella pinodes (Skiba et al., 2005), 89 nucleotide sequences mainly from Bowman–Birk protease inhibitor (BBI) coding sequences (41 accessions) and chloroplast sequences (21 accessions) and 216 protein sequences (44 amino acid sequences from BBI, 150 sequences from chloroplast proteins), for L. cicera these numbers were reduced to 4 internal transcribed spacers (ITS), 1 antifungal protein DNA sequence, 1 convicilin gene sequence, 26 sequences from chloroplast regions and 4 protein amino-acid sequences.

In order to perform precision plant breeding through marker assisted selection (MAS), it is necessary to identify the genetic regions that are closely linked to the genetic control of a particular trait of

17 Chapter 1 ______interest. Once an interesting plant trait is found associated with a marker (or more), plants can be selected, using a genetic screen with those markers, early on its growth stage. This selection allows a faster and more efficient breeding process.

Linkage maps are a representation of the relative position of genetic regions in the genome, taking into account the recombination frequency of those genetic regions in a segregating mapping population. Until now only two linkage maps using molecular markers were developed for L. sativus. One developed by Chowdhury and Slinkard (1999), using eleven Random Amplified Polymorphic DNA (RAPD) markers, one isozyme marker and one morphological trait (flower colour). The other linkage map was constructed by Skiba et al. (2004a), using 47 RAPDs, 7 cross- amplified pea microsatellite (SSR) markers and 13 Cleaved Amplified Polymorphic Sequence (CAPS) markers and was used to study the genetic basis of resistance to ascochyta blight. Nevertheless, these maps were not informative enough to allow bridging the information between them, as reviewed by Vaz Patto et al. (2006b).

Existing molecular markers specific or cross-amplification studies in grass pea included the work of Shiferaw et al. (2011), that successfully amplified nine EST-SSRs (expressed sequence tag - simple sequence repeats) developed from the EST sequences of Skiba et al. (2005) and 12 EST-SSRs from M. truncatula, which have been previously proven to be transferable to other legume species by Gutierrez et al. (2005). Lioi et al. (2011) were able to genotype in a grass pea diversity study, 10 SSRs developed from nucleotide sequences stored at public databases, being nine from L. sativus sequences and one from a L. japonicus sequence.

18 General introduction ______

Plant response to pathogens consists on the activation of several layers of defence, in a constant arms race between host and pathogen (Wirthmueller et al., 2013). After a compatible interaction, general defence mechanisms consist in perception through a panoply of receptors (Helliwell and Yang, 2001), that will mediate the expression of genes involved in hormone signalling, like the salicylic acid, jasmonic acid and ethylene pathways (Bari and Jones, 2009), leading to the reinforcement of plant cell wall through the production of callose or lignin, and the production of antimicrobial compounds in order to restrain pathogen development (Glazebrook, 2005).

Plants respond differentially to biotroph or necrotroph attack (Glazebrook, 2005). The most effective defence mechanism against biotrophic pathogens is the programmed cell death, preventing the pathogen from colonize adjacent host cells. On the other hand, necrotrophic pathogens feed from the debris of plant cells, and then benefit from the activation of the host cell death (Mengiste, 2012). Therefore, an efficient defence response of the plant relies on a dynamic recognition mechanism in order to prevent pathogen colonization.

The development of new molecular tools will allow the identification of candidate genes acting in the different phases of the host/pathogen interaction, increasing the knowledge on the defence mechanisms of Lathyrus spp.. Prior to the inclusion of this thesis results the only grass pea expression analysis existing was performed by Skiba et al. (2005), identifying 29 potential defence related genes differentially expressed in response to M. pinodes inoculation. These included genes associated with pathogen recognition, the phenylpropanoid pathway, hypersensitivity, pathogenesis-related and disease resistance response proteins.

19 Chapter 1 ______

1.10. Objectives

This thesis applies new technological advances in genetics and genomics to explore the genetics and mechanisms underlying Lathyrus spp. resistance to different pathogens. We have also developed new molecular tools to support future breeding efforts in these species.

In chapter 2, we aimed at evaluating the transferability of molecular markers developed for close related legume species to Lathyrus spp. and test the application of those new molecular tools on Lathyrus mapping and diversity analysis.

In chapter 3 and 5, efforts were made to unveil the different molecular responses, by a transcriptomic approach, of susceptible and resistant phenotypes of L. sativus (Chapter 3) and L. cicera (Chapter 5) to rust inoculation, and to develop new SSR and single nucleotide polymorphism (SNP) molecular markers from the RNA-Seq data generated.

In chapter 4, we aimed to elucidate the molecular responses, using deepSuperSAGE, of a resistant L. sativus genotype upon inoculation with Ascochyta lathyri.

In chapter 6, the objective was to develop the first linkage map for L. cicera and perform QTL analysis for resistance to rust and powdery mildew in this species and the first co-linearity studies with the model Medicago truncatula.

Finally, in chapter 7 it is discussed how the obtained results allowed the data integration in order to develop new molecular tools for Lathyrus ssp., providing a highly valuable resource for future smart breeding approaches in these previously under-researched, valuable legume crops.

20 General introduction ______

1.11. Acknowledgments

Part of this chapter is included in the book chapter “Grass pea”, Grain Legumes, Series Handbook of Plant Breeding. (New York, U.S.A.: Springer Science+Business Media) where NFA drafted the manuscript and DR and MCVP revised the manuscript critically.

1.12. References Abd El-Moneim, A.M., Van Dorrestein, B., Baum, M., and Mulugeta, W. (2000). Improving the nutritional quality and yield potential of grasspea (Lathyrus sativus L.). Food & Nutrition Bulletin 21, 493-496. Abd El Moneim, A., Van Dorrestein, B., Baum, M., Ryan, J., and Vejiga, G. (2001). Role of ICARDA in improving the nutritional quality and yield potential of grasspea (Lathyrus sativus L.) for subsistence farmers in dry areas. Lathyrus Lathyrism Newsletter 2, 55-58. Agrios, G.N. (2005). Plant Pathology. New York, U.S.A.: Elsevier Academic Press. Asmussen, C.B., and Liston, A. (1998). Chloroplast DNA characters, phylogeny, and classification of Lathyrus (Fabaceae). American Journal of Botany 85, 387-401. Bari, R., and Jones, J.G. (2009). Role of plant hormones in plant defence responses. Plant Molecular Biology 69, 473-488. Barik, D.P., Mohapatra, U., and Chand, P.K. (2005). Transgenic grasspea (Lathyrus sativus L.): factors influencing agrobacterium-mediated transformation and regeneration. Plant Cell Reports 24, 523-531. Barilli, E., Moral, A., Sillero, J.C., and Rubiales, D. (2012). Clarification on rust species potentially infecting pea (Pisum sativum L.) crop and host range of Uromyces pisi (Pers.) Wint. Crop Protection 37, 65-70. Barilli, E., Satovic, Z., Sillero, J.C., Rubiales, D., and Torres, A.M. (2011). Phylogenetic analysis of Uromyces species infecting grain and forage legumes by sequence analysis of nuclear ribosomal internal transcribed spacer region. Journal of Phytopathology 159, 137-145. Ben Brahim, N., Combes, D., and Marrakchi, M. (2001). Autogamy and allogamy in genus Lathyrus. Lathyrus Lathyrism Newsletter 2, 21-26.

21 Chapter 1 ______

Biswas, A.K. (2007). "Induced mutation in grass pea," in Breeding of neglected and under-utilised crops, spices and herbs., eds. S.J. Ochatt & S.M. Jain. (Plymouth, U.K.: Science Publishers), 29- 39. Campbell, C.G. (1997). Promoting the conservation and use of underutilized and neglected crops. Grass pea. Lathyrus sativus L. Rome, Italy: International Plant Genetic Resources Institute. Campbell, C.G., and Briggs, C.J. (1987). Registration of low neurotoxin content Lathyrus germplasm LS 8246. Crop Science 27, 821- 821. Campbell, C.G., Mehra, R.B., Agrawal, S.K., Chen, Y.Z., Elmoneim, A.M.A., Khawaja, H.I.T., Yadov, C.R., Tay, J.U., and Araya, W.A. (1994). Current status and future strategy in breeding grasspea (Lathyrus sativus). Euphytica 73, 167-175. Carita, T. (2012). Influencia de la fecha de siembra, genotipo y densidad de plantas en el crecimiento, rendimiento y calidad del Lathyrus sativus L. en condiciones de secano Mediterráneo. PhD, UNIVERSIDAD DE EXTREMADURA. Carrillo, E., Rubiales, D., Perez-De-Luque, A., and Fondevilla, S. (2013). Characterization of mechanisms of resistance against Didymella pinodes in Pisum spp. European Journal of Plant Pathology 135, 761-769. Chowdhury, M.A., and Slinkard, A.E. (1997). Natural outcrossing in grasspea. Journal of Heredity 88, 154-156. Chowdhury, M.A., and Slinkard, A.E. (1999). Linkage of random amplified polymorphic DNA, isozyme and morphological markers in grasspea (Lathyrus sativus). Journal of Agricultural Science 133, 389-395. Chowdhury, M.A., and Slinkard, A.E. (2000). Genetic diversity in grasspea (Lathyrus sativus L.). Genetic Resources and Crop Evolution 47, 163-169. Duke, J. (1981). Handbook of Legumes of World Economic Importance. New York, U.S.A.: Plennun Press. Ellison, N.W., Liston, A., Steiner, J.J., Williams, W.M., and Taylor, N.L. (2006). Molecular phylogenetics of the clover genus (Trifolium- Leguminosae). Molecular Phylogenetics and Evolution 39, 688- 705. Erskine, W., Smartt, J., and Muehlbauer, F. (1994). Mimicry of lentil and the domestication of common vetch and grass pea. Economic Botany 48, 326-332. Fao (2009). International treaty on plant genetic resources for food and agriculture, Food and Agriculture Organization of the United Nations (FAO). Rome, Italy: FAO (accessible at ftp://ftp.fao.org/docrep/fao/011/i0510e/i0510e.pdf).

22 General introduction ______

Fernández-Aparicio, M., Flores, F., and Rubiales, D. (2009). Field response of Lathyrus cicera germplasm to crenate broomrape (Orobanche crenata). Field Crops Research 113, 321-327. Fernández-Aparicio, M., Flores, F., and Rubiales, D. (2012). Escape and true resistance to crenate broomrape (Orobanche crenata Forsk.) in grass pea (Lathyrus sativus L.) germplasm. Field Crops Research 125, 92-97. Fernández-Aparicio, M., and Rubiales, D. (2010). Characterisation of resistance to crenate broomrape (Orobanche crenata Forsk.) in Lathyrus cicera L. Euphytica 173, 77-84. Fikre, A., Korbu, L., Kuo, Y.H., and Lambein, F. (2008). The contents of the neuro-excitatory amino acid beta-ODAP (beta-N-oxalyl- L-alpha,beta-diaminopropionic acid), and other free and protein amino acids in the seeds of different genotypes of grass pea (Lathyrus sativus L.). Food Chemistry 110, 422-427. Gcdt (2009). Strategy for the ex situ conservation of Lathyrus (grass pea), with special reference to Lathyrus sativus, L. cicera, and L. ochrus. Rome, Italy: Global Crop Diversity Trust (GCDT). Glazebrook, J. (2005). Contrasting mechanisms of defense against biotrophic and necrotrophic pathogens. Annual Review of Phytopathology 43, 205-227. Granati, E., Bisignano, V., Chiaretti, D., Crinò, P., and Polignano, G. (2003). Characterization of Italian and exotic Lathyrus germplasm for quality traits. Genetic Resources and Crop Evolution 50, 273-280. Grela, E., Rybiński, W., Matras, J., and Sobolewska, S. (2012). Variability of phenotypic and morphological characteristics of some Lathyrus sativus L. and Lathyrus cicera L. accessions and nutritional traits of their seeds. Genetic Resources and Crop Evolution 59, 1687-1703. Gurung, A.M., and Pang, E.C.K. (2011). "Lathyrus," in Wild Crop Relatives: Genomic and Breeding Resources, ed. C. Kole. (Berlin, Germany: Springer-Verlag), 117-126. Gurung, A.M., Pang, E.C.K., and Taylor, P.W.J. (2002). Examination of Pisum and Lathyrus species as sources of ascochyta blight resistance for field pea (Pisum sativum). Australasian Plant Pathology 31, 41-45. Gutierrez, M.V., Vaz Patto, M.C., Huguet, T., Cubero, J.I., Moreno, M.T., and Torres, A.M. (2005). Cross-species amplification of Medicago truncatula microsatellites across three major pulse crops. Theoretical and Applied Genetics 110, 1210-1217. Hanbury, C.D., Sarker, A., Siddique, K.H.M., and Perry, M.W. (1995). Evaluation of Lathyrus germplasm in a Mediterranean type environment in south-western Australia. CLIMA Occasional

23 Chapter 1 ______

Publication (Cooperative Research Centre for Legumes in Mediterranean Agriculture) No. 8, Perth, Australia, p 23. Hanbury, C.D., Siddique, K.H.M., Galwey, N.W., and Cocks, P.S. (1999). Genotype-environment interaction for seed yield and ODAP concentration of Lathyrus sativus L. and L. cicera L. in Mediterranean-type environments. Euphytica 110, 45-60. Hanbury, C.D., White, C.L., Mullan, B.P., and Siddique, K.H.M. (2000). A review of the potential of Lathyrus sativus L. and L. cicera L. grain for use as animal feed. Animal Feed Science and Technology 87, 1-27. Hanbury, C.L., and Siddique, K.H.M. (2000). Registration of 'Chalus' Lathyrus cicera L. Crop Science 40, 1199. Helliwell, E.E., and Yang, Y. (2001). "Systemic signalling in plant defence," in eLS. (Chichester, U.K.: John Wiley & Sons, Ltd). Heywood, V., Casas, A., Ford-Lloyd, B., Kell, S., and Maxted, N. (2007). Conservation and sustainable use of crop wild relatives. Agriculture Ecosystems & Environment 121, 245-255. Hillocks, R., and Maruthi, M. (2012). Grass pea (Lathyrus sativus): Is there a case for further crop improvement? Euphytica 186, 647- 654. Hopf, M. (1986). "Archaeological evidence of the spread and use of some members of the Leguminosae family," in The Origin and domestication of cultivated plants: symposium, ed. C. Barigozzi. (Oxford, U.K.: Elsevier), 35-60. Icar (2009). Project Coordinator’s Report of All India Coordinated Research Project on Mungbean, Urdbean, Lentil, Lathyrus, Rajmash and Pea. New Delhi, India: Indian Council of Agricultural Research, pp.118. Icarda (2006). ICARDA Annual Report 2005. Aleppo, Syria: International Center for Agricultural Research in the Dry Areas. pp. 54-55. Icarda (2007). ICARDA Annual Report 2006. Aleppo, Syria: International Center for Agricultural Research in the Dry Areas. pp. 57-58. Jackson, M.T., and Yunus, A.G. (1984). Variation in the grass pea (Lathyrus sativus L.) and Wild-Species. Euphytica 33, 549-559. Kenicer, G., Nieto-Blásquez, E.M., Mikić, A., and Smýkal, P. (2009). Lathyrus – diversity and phylogeny in the genus. Grain Legumes 54, 16-18. Kenicer, G.J., Kajita, T., Pennington, R.T., and Murata, J. (2005). Systematics and biogeography of Lathyrus (Leguminosae) based on internal transcribed spacer and cpDNA sequence data. American Journal of Botany 92, 1199-1209.

24 General introduction ______

Kislev, M.E. (1989). Origins of the Cultivation of Lathyrus sativus and Lathyrus cicera (Fabaceae). Economic Botany 43, 262-270. Kiyoshi, Y., Toshiyuki, F., and Blumenreich, I.D. (1985). "Isozymic variation and interspecific crossability in annual species of genus Lathyrus L.," in Lathyrus and Lathyrism, eds. A.K. Kaul & D. Combes. (Pau, France), 118-129. Krause, D., and Krause, I. (2003). New green manuring Lathyrus sativus variety AC Greenfix available in USA. Lathyrus Lathyrism Newsletter 3, 13-14. Kumar, S., Bejiga, G., Ahmed, S., Nakkoul, H., and Sarker, A. (2011). Genetic improvement of grass pea for low neurotoxin (beta- ODAP) content. Food and Chemical Toxicology 49, 589-600. Kumar, S., Gupta, P., Barpete, S., Sarker, A., Amri, A., Mathur, P.N., and Baum, M. (2013). "Grass Pea," in Genetic and Genomic Resources of Grain Legume Improvement, eds. M. Singh, H.D. Upadhyaya & I.S. Bisht. (Oxford, U.K.: Elsevier Science), 269- 292. Kuo, Y.H., Bau, H.M., Rozan, P., Chowdhury, B., and Lambein, F. (2000). Reduction efficiency of the neurotoxin beta-ODAP in low-toxin varieties of Lathyrus sativus seeds by solid state fermentation with Aspergillus oryzae and Rhizopus microsporus var. chinensis. Journal of the Science of Food and Agriculture 80, 2209-2215. Kupicha, F.K. (1983). The infrageneric structure of Lathyrus. Notes from the Royal Botanic Garden Edinburgh 41, 209-244. Lal, M.S., Agrawal, I., and Chitale, M.W. (1986). "Genetic improvement of chickling vetch (Lathyrus sativus L.) in Madhya Pradesh, India," in Lathyrus and Lathyrism, eds. A.K. Kaul & D. Combes. (New York, U.S.A.: Third World Medical Research Foundation), 146-160. Lambein, F., Ngudi, D.D., and Kuo, Y.H. (2009). Progress in prevention of toxico-nutritional neurodegenerations. http:// www.atdforum.org/journal/html/200 9-34/8. Accessed July 2014. Lioi, L., Sparvoli, F., Sonnante, G., Laghetti, G., Lupo, F., and Zaccardelli, M. (2011). Characterization of Italian grasspea (Lathyrus sativus L.) germplasm using agronomic traits, biochemical and molecular markers. Genetic Resources and Crop Evolution 58, 425-437. Malek, M.A. (1998). "Genetic resources of grass pea (Lathyrus sativus L.) in Bangladesh. Proceedings of the IPGRIICARDA-ICAR Regional Working Group Meeting", in: Lathyrus Genetic Resources Network. (eds.) P.N. Mathur, V.R. Rao & A. R.K.

25 Chapter 1 ______

(New Delhi, India: National Bureau of Plant Genetic Resources, IPGRI). Martín-Sanz, A., Palomo, J., Pérez De La Vega, M., and Caminero, C. (2012). Characterization of Pseudomonas syringae pv. syringae isolates associated with bacterial blight in Lathyrus spp. and sources of resistance. European Journal of Plant Pathology 134, 205-216. Mengiste, T. (2012). Plant immunity to necrotrophs. Annual Review of Phytopathology 50, 267-294. Mera, M. (2010). Inheritance of seed weight in large-seed grass pea Lathyrus sativus L. Chilean Journal of Agricultural Research 70, 357-364. Mera, M., Tay, J., France, A., Montenegro, A., Espinoza, A., and Guete, N. (2003). Luanco-INIA, a large-seeded cultivar of Lathyrus sativus released in Chile. Lathyrus Lathyrism Newsletter 3, 26. Nerkar, Y.S. (1972). Induced variation and response to selection for low neurotoxin content in Lathyrus sativus. Indian Journal of Genetics and Plant Breeding 32, 175-180. Nerkar, Y.S. (1976). Mutation studies in Lathyrus sativus. Indian Journal of Genetics and Plant Breeding 36, 223-229. O’connell, R.J., and Panstruga, R. (2006). Tête à tête inside a plant cell: establishing compatibility between plants and biotrophic fungi and oomycetes. New Phytologist 171, 699-718. Ochatt, S.J., Conreux, C., and Jacas, L. (2013). Flow cytometry distinction between species and between landraces within Lathyrus species and assessment of true-to-typeness of in vitro regenerants. Plant Systematics and Evolution 299, 75-85. Ochatt, S.J., Muneaux, E., Machado, C., Jacas, L., and Pontécaille, C. (2002a). The hyperhydricity of in vitro regenerants of grass pea (Lathyrus sativus L.) is linked with an abnormal DNA content. Journal of Plant Physiology 159, 1021-1028. Ochatt, S.J., Sangwan, R.S., Marget, P., Ndong, Y.A., Rancillac, M., Perney, P., and Röbbelen, G. (2002b). New approaches towards the shortening of generation cycles for faster breeding of protein legumes. Plant Breeding 121, 436-440. Pandey, R.L., Chitale, M.W., Sharma, R.N., and Rastogi, N. (1996). "Status of Lathyrus research in India. Proceedings of a Regional Workshop", in: Lathyrus Genetic Resources in Asia. (eds.) R.K. Arora, P.N. Mathur, K.W. Riley & Y. Adham. (New Delhi, India: International Plant Genetic Resourses Institute). Parlevliet, J.E. (1979). "The co-evolution of host-parasite systems," in Parasites as plant taxonomists, Acta Universitatis Upsaliensis XXVI, ed. J. Hedberg.), 39-45.

26 General introduction ______

Peever, T.L. (2007). Role of host specificity in the speciation of Ascochyta pathogens of cool season food legumes. European Journal of Plant Pathology 119, 119-126. Peña-Chocarro, L., and Peña, L.Z. (1999). History and traditional cultivation of Lathyrus sativus L. and Lathyrus cicera L. in the Iberian peninsula. Vegetation History and Archaeobotany 8, 49- 52. Rahman, M.M., Kumar, J., Rahman, M.A., and Ali Afzal, M. (1995). Natural outcrossing in Lathyrus sativus L. Indian Journal of Genetics & Plant Breeding 55, 204-207. Roy, P.K., All, K., Gupta, A., Barat, G.K., and Mehta, S.L. (1993). β-N- Oxalyl-L-α, β-diaminopropionic acid in somaclones derived from internode explants of Lathyrus sativus. Journal of Plant Biochemistry and Biotechnology 2, 9-13. Rubiales, D., and Fondevilla, S. (2012). Future prospects for ascochyta blight resistance breeding in cool season food legumes. Frontiers in Plant Science 3, 27. Rubiales, D., Sillero, J.C., and Emeran, A.A. (2013). Response of vetches (Vicia spp.) to specialized forms of Uromyces vicia- fabae and to Uromyces pisi. Crop Protection 46, 38-43. Rybinski, W. (2003). Mutagenesis as a tool for improvement of traits in grass pea (Lathyrus sativus L.). Lathyrus Lathyrism Newsletter 3, 30-34. Schaefer, H., Hechenleitner, P., Santos-Guerra, A., Sequeira, M.M.D., Pennington, R.T., Kenicer, G., and Carine, M.A. (2012). Systematics, biogeography, and character evolution of the legume tribe Fabeae with special focus on the middle-Atlantic island lineages. BMC Evolutionary Biology 12, 250. Shiferaw, E., Pè, M., Porceddu, E., and Ponnaiah, M. (2011). Exploring the genetic diversity of Ethiopian grass pea (Lathyrus sativus L.) using EST-SSR markers. Molecular Breeding 30, 789-797. Siddique, K.H.M., Hanbury, C.L., and Sarker, A. (2006). Registration of ‘Ceora’ grass pea registration by CSSA. Crop Science 46, 986-986. Sillero, J.C., Cubero, J.I., Fernández-Aparicio, M., and Rubiales, D. (2005). Search for resistance to crenate broomrape (Orobanche crenata) in Lathyrus. Lathyrus Lathyrism Newsletter 4, 7-9. Sillero, J.C., Fondevilla, S., Davidson, J., Patto, M.C.V., Warkentin, T.D., Thomas, J., and Rubiales, D. (2006). Screening techniques and sources of resistance to rusts and mildews in grain legumes. Euphytica 147, 255-272. Skiba, B., Ford, R., and Pang, E.C.K. (2004a). Construction of a linkage map based on a Lathyrus sativus backcross population

27 Chapter 1 ______

and preliminary investigation of QTLs associated with resistance to ascochyta blight. Theoretical and Applied Genetics 109, 1726-1735. Skiba, B., Ford, R., and Pang, E.C.K. (2004b). Genetics of resistance to Mycosphaerella pinodes in Lathyrus sativus. Australian Journal of Agricultural Research 55, 953–960. Skiba, B., Ford, R., and Pang, E.C.K. (2005). Construction of a cDNA library of Lathyrus sativus inoculated with Mycosphaerella pinodes and the expression of potential defence-related expressed sequence tags (ESTs). Physiological and Molecular Plant Pathology 66, 55-67. Smartt, J. (1984). Evolution of grain legumes .1. Mediterranean pulses. Experimental Agriculture 20, 275-296. Smith, P.H., Foster, E.M., Boyd, L.A., and Brown, J.K.M. (1996). The early development of Erysiphe pisi on Pisum sativum L. Plant Pathology 45, 302-309. Smýkal, P., Kenicer, G., Flavell, A.J., Corander, J., Kosterin, O., Redden, R.J., Ford, R., Coyne, C.J., Maxted, N., Ambrose, M.J., and Ellis, N.T.H. (2011). Phylogeny, phylogeography and genetic diversity of the Pisum genus. Plant Genetic Resources 9, 4-18. Staples, R.C. (2000). Research on the rust fungi during the twentieth century. Annual Review of Phytopathology 38, 49-69. Tadesse, W., and Bekele, E. (2003). Phenotypic diversity of Ethiopian grass pea (Lathyrus sativus L.) in relation to geographical regions and altitudinal range. Genetic Resources and Crop Evolution 50, 497-505. Talukdar, D. (2009). Dwarf mutations in grass pea (Lathyrus sativus L.): origin, morphology, inheritance and linkage studies. Journal of Genetics 88, 165-175. Talukdar, D. (2011). Isolation and characterization of NaCl-tolerant mutations in two important legumes, Clitoria ternatea L. and Lathyrus sativus L.: Induced mutagenesis and selection by salt stress. Journal of Medicinal Plants Research 5, 3619-3628. Tivoli, B., and Banniza, S. (2007). Comparison of the epidemiology of ascochyta blights on grain legumes. European Journal of Plant Pathology 119, 59-76. Tivoli, B., Baranger, A., Avila, C.M., Banniza, S., Barbetti, M., Chen, W.D., Davidson, J., Lindeck, K., Kharrat, M., Rubiales, D., Sadiki, M., Sillero, J.C., Sweetingham, M., and Muehlbauer, F.J. (2006). Screening techniques and sources of resistance to foliar diseases caused by major necrotrophic fungi in grain legumes. Euphytica 147, 223-253.

28 General introduction ______

Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2006a). Characterization of resistance to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus sativus. Plant Breeding 125, 308-310. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2007). Resistance reaction to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus cicera from Iberian origin. Genetic Resources and Crop Evolution 54, 1517-1521. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2009). Pre and posthaustorial resistance to rusts in Lathyrus cicera L. Euphytica 165, 27-34. Vaz Patto, M.C., Hanbury, C.D., Moorhem, M.V., Lambein, F., Ochatt, S.J., and Rubiales, D. (2011). "Grass Pea," in Genetics, Genomics and Breeding of Cool Season Grain Legumes, ed. C. Kole. (Enfield, U.S.A.: Science Publishers), 151-204. Vaz Patto, M.C., and Rubiales, D. (2009). Identification and characterization of partial resistance to rust in a germplasm collection of Lathyrus sativus L. Plant Breeding 128, 495-500. Vaz Patto, M.C., and Rubiales, D. (2014a). Lathyrus diversity: available resources with relevance to crop improvement – L. sativus and L. cicera as case studies. Annals of Botany 113, 895-908. Vaz Patto, M.C., and Rubiales, D. (2014b). Resistance to rust and powdery mildew in Lathyrus crops. Czech Journal of Genetics and Plant Breeding 50, 116-122. Vaz Patto, M.C., Skiba, B., Pang, E.C.K., Ochatt, S.J., Lambein, F., and Rubiales, D. (2006b). Lathyrus improvement for resistance against biotic and abiotic stresses: From classical breeding to marker assisted selection. Euphytica 147, 133-147. Wirthmueller, L., Maqbool, A., and Banfield, M.J. (2013). On the front line: structural insights into plant-pathogen interactions. Nature Reviews Microbiology 11, 761-776. Wojciechowski, M.F., Lavin, M., and Sanderson, M.J. (2004). A phylogeny of legumes (Leguminosae) based on analysis of the plastid matK gene resolves many well-supported subclades within the family. American Journal of Botany 91, 1846-1862. Yadav, C.R. (1996). "Genetic evaluation and varietal improvement of grasspea in Nepal," in Lathyrus Genetic Resources in Asia: Proc Regional Workshop. Indira Gandhi Agricultural University, Raipur, India. IPGRI Office for South Asia, New Delhi, India, pp 21-28., eds. R.K. Arora, P.N. Mathur, K.W. Riley & Y. Adham.). Yang, H.-M., and Zhang, X.-Y. (2005). Considerations on the reintroduction of grass pea in China. Lathyrus Lathyrism Newsletter 4, 22-26.

29 Chapter 1 ______

Yunus, A.G., and Jackson, M.T. (1991). The gene pools of the grasspea (Lathyrus sativus L.). Plant Breeding 106, 319-328. Zambre, M., Chowdhury, B., Kuo, Y.-H., Van Montagu, M., Angenon, G., and Lambein, F. (2002). Prolific regeneration of fertile plants from green nodular callus induced from meristematic tissues in Lathyrus sativus L. (grass pea). Plant Science 163, 1107-1112.

30 CHAPTER 2

Transferability of molecular markers from major legumes to Lathyrus spp. for their application in mapping and diversity studies

The work presented in this chapter was mostly performed by Nuno Felipe Almeida (see Acknowledgements section), and corresponds to the following publication:

Almeida, N.F., Leitão, S.T., Caminero, C., Torres, A.M., Rubiales, D., and Vaz Patto, M.C. (2014). Transferability of molecular markers from major legumes to Lathyrus spp. for their application in mapping and diversity studies. Molecular Biology Reports 41, 269-283.

31

Cross-species amplification ______

2.1. Abstract

Lathyrus cicera L. (chickling pea) and L. sativus L. (grass pea) have great potential among grain legumes due to their adaptability to inauspicious environments, high protein content and resistance to serious diseases. Nevertheless, due to its past underused, further activities are required to exploit this potential and to capitalise on the advances in molecular biology that enable improved Lathyrus spp. breeding programmes. In this study we evaluated the transferability of molecular markers developed for closely related legume species to Lathyrus spp. (Medicago truncatula, pea, lentil, fababean and lupin) and tested the application of those new molecular tools on Lathyrus mapping and diversity studies. Genomic and expressed sequence tag microsatellite (gSSR and EST-SSR), intron-targeted amplified polymorphic (ITAP), resistance gene analogue (RGA) and defence- related gene (DR) markers were tested. In total 128 (27.7%) and 132 (28.6%) molecular markers were successfully cross-amplified, respectively in L. cicera and L. sativus. In total, the efficiency of transferability from genomic microsatellites was 5%, and from gene- based markers, 55%. For L. cicera, three Cleaved Amplified Polymorphic Sequence markers (CAPS) and one derived Cleaved Amplified Polymorphic Sequence marker (dCAPS) based on the cross-amplified markers were also developed. Nine of those molecular markers were suitable for mapping in a L. cicera Recombinant Inbred Line (RIL) population. From the 17 molecular markers tested for diversity analysis, six (35%) in L. cicera and seven (41%) in L. sativus were polymorphic and discriminate well all the L. sativus accessions. Additionally, L. cicera accessions were clearly distinguished from L. sativus accessions. This work revealed a high number of transferable molecular markers to be used in current

33 Chapter 2 ______genomic studies in Lathyrus spp.. Although their usefulness was higher on diversity studies, they represent the first steps for future comparative mapping involving these species.

2.2. Introduction

Lathyrus sativus L. (grass pea) and L. cicera L. (chickling pea) are legume crops with considerable potential in dryland farming systems of semi-arid regions. Their ability to provide an economic yield under adverse conditions made them popular crops in subsistence farming in many developing countries, offering great potential for use in marginal low-rainfall areas (Campbell, 1997). Although widespread in the past, both as forage and grain crops, they are now rarely grown in Europe due to yield unpredictability and the presence of anti-nutritional substances. However, a renewed interest in its reintroduction in cropping systems in Southern Australia and North America is growing because of their high agronomic potential (Hanbury et al., 2000; Rao and Northup, 2011; Calderón et al., 2012; Gusmao et al., 2012). In Europe, their cultivation is justified by the need to recover marginal lands, providing also an efficient alternative to the areas overexploited by cereal cultivation (Vaz Patto et al., 2006b; Tavoletti and Iommarini, 2007; Grela et al., 2012; Martín-Sanz et al., 2012). Moreover, a large variation has been found in Lathyrus spp. germplasm regarding the resistance for common diseases in grain legumes (Vaz Patto et al., 2006a; 2007; 2009; Vaz Patto and Rubiales, 2009).

L. sativus and L. cicera are diploid (2n = 14) and primarily self- pollinated (Ben Brahim et al., 2001). Due to Lathyrus large genome size, circa 8.2 Gb in L. sativus and circa 6.8 Gb in L. cicera (reviewed

34 Cross-species amplification ______by Bennett and Leitch, 2012) and little economic relevance in developed countries, not much progress has been attained on the study of the genetic control of important traits such as disease resistance, hampering the development of modern cultivars or the introgression of their interesting traits in other related species. Until now only two linkage maps using molecular markers were developed for L. sativus (Chowdhury and Slinkard, 1999; Skiba et al., 2004), but these maps were not informative enough to bridge that information between both of them, as reviewed by Vaz Patto et al. (2006b).

Cross-species and cross-genus amplification of molecular markers is now a common strategy for the discovery of markers to use on the not so well studied species (Castillo et al., 2008; Ellwood et al., 2008). Molecular markers from genomic libraries (genomic microsatellites, gSSRs) and/or derived from expressed sequence tags from the most important crops or model species, are now frequently used on diversity, evolutionary and mapping studies in other related species (Xu et al., 2004; Zhang et al., 2007; Feng et al., 2009; Datta et al., 2010; Harris-Shultz et al., 2012). These valuable tools can also be used for comparative mapping between underused food and feed legume crops and the model species. Within the Fabaceae family, this relationship can work both ways (Varshney et al., 2009; Parra-Gonzalez et al., 2012) for the “orphan” species, well studied and phylogenetically close species can provide molecular tools for genetic and genomic studies; and in the opposite direction, underused crops such as the Lathyrus spp. can be a source of interesting genes such as biotic and abiotic resistance. From the better genetically characterized legume species (Kumar et al., 2012), Pisum sativum L. is reported as the closest related to Lathyrus, followed by lentil, faba bean, Medicago, chickpea and Lotus

35 Chapter 2 ______

(Asmussen and Liston, 1998; Wojciechowski et al., 2004; Smýkal et al., 2011; Schaefer et al., 2012).

In this study we evaluated the transferability of molecular markers developed for Medicago truncatula Gaertn., P. sativum (pea), Lens culinaris Medik. (lentil), Lupinus spp. and Vicia faba L. (faba bean), to Lathyrus spp. and tested the application of those new molecular tools on Lathyrus mapping and diversity studies. This represents the first steps for future comparative mapping in Lathyrus spp..

2.3. Methods

2.3.1. Plant Material and DNA isolation

For the cross-amplification screening two L. sativus accessions (BGE015746, BGE024709) and two L. cicera accessions (BGE008277, BGE023542), kindly provided by the Plant Genetic Resources Centre (CRF-INIA), Madrid, Spain were used. These accessions were already evaluated for resistance against rust and powdery mildew, and showed, within each particular species, contrasting phenotypes (Vaz Patto et al., 2006a; 2007; 2009; Vaz Patto and Rubiales, 2009). Cross-amplification controls used were: the P. sativum cv. ‘Messire’ and the P. sativum subsp. syriacum accession P665, parents of a recombinant inbred line RILs mapping population (Fondevilla et al., 2008; Fondevilla et al., 2010; Fondevilla et al., 2011), the L. culinaris cv. ‘Armuña’, the M. truncatula cv. ‘Jemalong’ and the V. faba accession Vf136. To validate the usefulness of transferable markers for diversity studies, 20 accessions randomly chosen of L. sativus and L. cicera (10 of each) from the CRF-INIA collection were used (Table 2.1). To validate

36 Cross-species amplification ______usefulness of the transferable markers in the development of a linkage map, 103 individuals from a L. cicera RILs F5 population, segregating for several fungal diseases resistance, were used. DNA from fresh young leaves was extracted using a modified CTAB protocol developed by Torres et al. (1993).

Table 2.1 – Lathyrus cicera and Lathyrus sativus accession references at the CRF-INIA germplasm bank. For more information consult: http://wwwx.inia.es/crf/WWWCRF/CRFing/PaginaPrincipal.asp

CRF-INIA Collection site Species accession Latitude Longitude Altitude (m) BGE001043 374329N 0035758W 753 Lathyrus cicera L. BGE001164 385138N 0060555W 287 Lathyrus cicera L. BGE008277 385633N 0031416W 770 Lathyrus cicera L. BGE010898 424901N 0042242W 991 Lathyrus cicera L. BGE018818 395520N 0025259W 870 Lathyrus cicera L. BGE022223 373857N 0020422W 839 Lathyrus cicera L. BGE023542 373149N 0033908W 1084 Lathyrus cicera L. BGE023558 - - - Lathyrus cicera L. BGE023562 - - - Lathyrus cicera L. BGE027064 365251N 0031752W 684 Lathyrus cicera L. BGE001489 385403N 0030311W 863 Lathyrus sativus L. BGE002259 420039N 0060233W 750 Lathyrus sativus L. BGE003490 420039N 0060233W 750 Lathyrus sativus L. BGE015744 384210N 002361W 870 Lathyrus sativus L. BGE015746 394454N 0015846W 991 Lathyrus sativus L. BGE018800 400835N 0013801W 1237 Lathyrus sativus L. BGE023247 424425N 0063930W 670 Lathyrus sativus L. BGE023500 420032N 0043200W 734 Lathyrus sativus L. BGE023552 - - - Lathyrus sativus L. BGE023553 - - - Lathyrus sativus L. BGE024709 290935N 0132947W 241 Lathyrus sativus L. BGE029748 400833N 0032516W 753 Lathyrus sativus L.

37 Chapter 2 ______

2.3.2. Microsatellite markers

Two hundred and twenty seven simple sequence repeat (SSRs) primer pairs (Loridon et al., 2005), from which 42 expressed sequence tag microsatellite (EST-SSR) markers (Burstin et al., 2001) and 185 gSSR markers (Pea Microsatellite Consortium set up by Agrogène Inc., Moissy-Cramayel, France) developed for pea, 30 primer pairs of gSSR markers developed for lentil (Hamwieh et al., 2005) and 25 gSSR primer pairs developed for faba bean (Pozarkova et al., 2002) were tested (see ESM 2.1 for sequences and optimized annealing temperature (Ta)).

PCR reactions were conducted in a total volume of 15 μl using a Biometra Uno II thermal cycler, containing 20 ng of template DNA, 0.2 μM of forward primer and of reverse primer, 0.2 mM of each dNTP, 1.5-2 mM of MgCl2, and 0.6 units of Taq DNA polymerase (Promega, Madison, USA). The amplification reaction consisted of a denaturing step of 2 min at 95ºC, followed by 35 cycles of 30 s at

94ºC, 30 s at the optimized annealing temperature (Ta), and 30 s at

72ºC. For the pea EST-SSRs and gSSRs and lentil gSSRs Ta was optimized using as starting point the optimal temperature described for the donor species (see ESM 2.1). For the faba bean gSSRs Ta was optimized using 58ºC as initial temperature. The reaction was terminated at 72ºC for 5 min. SSR fragments were resolved using a 1.5 % Seaken LE TBE agarose gel (Lonza, Rockland, USA) with 0.5 μg/L ethidium bromide and visualised using a GEL-DOC 1000 System (Bio-Rad, Hercules, USA).

38 Cross-species amplification ______

2.3.3. Intron-Targeted Amplified Polymorphic markers, Defence- Related Genes and Resistance Gene Analogues

One hundred fifty six primers pairs of intron-targeted amplified polymorphic (ITAPs) markers from M. truncatula, P. sativum and Lupinus spp. were tested on the Lathyrus spp. lines (Phan et al., 2007; Ellwood et al., 2008). These markers were developed after sequence alignments of M. truncatula and Lupinus spp. (ITAP ML) (Nelson et al., 2010), M. truncatula, Lupinus albus and Glycine max (ITAP MLG) (Phan et al., 2007), M. truncatula and M. sativa (ITAP MP) (Choi et al., 2004). The ITAPs from the Grain Legume Integrated Program (ITAP GLIP) were developed mainly from sequence alignments of M. truncatula and pea as referred by Ellwood et al. (2008). Additionally, primer pairs designed for 12 defence-related (DR) genes and 12 (putative nucleotide-binding site leucine-rich repeat, NBS-LRR, type) resistance gene analogues (RGAs) developed for pea (Prioul-Gervais et al., 2007), were tested with the Lathyrus accessions previously described. A full list with all primers’ sequence, Ta and cited reference can be found in ESM 2.1. The molecular markers from the GLIP project are also available at the website (http://bioweb.abc.hu/cgi-mt/pisprim/pisprim.pl).

PCR reactions were conducted in a total volume of 20 µl, containing 30 ng of template DNA, 0.6 µM of forward primer and of reverse primer, 0.2 mM of each dNTP, 1.5-2 mM of MgCl2, and 1 U of Taq DNA polymerase (Biotools, Madrid, Spain). The amplification reaction consisted of a denaturing step of 5 min at 95ºC, followed by 40 cycles of 60 s at 95ºC, 60 s at 58-62ºC, and 2 min at 72ºC (see

ESM 2.1 for sequences and Ta). The reaction was terminated at 72ºC for 8 min. The amplified fragments were resolved using 2 % w/v (1 % w/v Seakem, 1 % w/v NuSieve) agarose gels in 1 x TBE buffer. Gels

39 Chapter 2 ______were stained with ethidium bromide and photographed under UV light using the software KODAK Digital Science 1D Ver. 2.0 and 3.5.

2.3.4. Sequencing

All primer combinations were amplified at least two times to rule out nonspecific amplicons or the possibility of PCR failure. Primer pairs that produced a complex amplification pattern or presented non consistent amplicon sizes after the second PCR reaction were not further analysed. Primers producing one band in Lathyrus spp. of similar length to the donor species were re-amplified and resolved in a 4 % Metaphor gel (Lonza, Rockland, USA), to confirm the presence of a single band.

PCR products that originated one confirmed single amplicon were purified using MultiScreen PCRµ96 Filter Plate, (Millipore, Billerica, USA) and sequenced in both directions using BigDye Terminator 3 on an ABI 3730XL sequencer (Applied Biosystems). Obtained sequences were aligned against the donor sequence, using the software Geneious (Drummond et al., 2011), to confirm the amplification of the same locus.

2.3.5. Segregation and diversity studies

To test the applicability of the cross amplified markers for mapping and diversity studies, an M13 tail was added to the 5’-end of the forward primers, allowing them to be labelled with IRD fluorescence (Schuelke, 2000), to allow resolution using a LI-COR 4300 DNA Analyzer (Lincoln, NE, USA). PCR reactions were conducted in a total volume of 10 µl containing 10 ng of template

40 Cross-species amplification ______

DNA, 0.04 µM of M13(-21) tagged forward primer, 0.16 µM of IRD700 or IRD 800 M13(-21) and 0.16 µM of reverse primer, 0.2 mM of each dNTP, 1.5 mM of MgCl2, and 0.2 unit of Taq DNA polymerase (Promega, Madison, USA). The amplification reaction consisted of a denaturing step of 5 min at 94ºC, followed by 30 cycles of 30 s at 94ºC, 45 s at 56ºC, 45 s at 72ºC, and 8 cycles of 30 s at 94ºC, 45 s at 53ºC, 45 s at 72ºC. The reaction was terminated at 72ºC for 10 min.

For the segregation study of each assessed marker, a chi- square analysis was used to test for deviations from the expected 1:1 segregation ratio in the RIL population. For the diversity study, statistics on diversity, including number of alleles per locus, major allele frequency, gene diversity, heterozygosity and Polymorphic Information Content (PIC) values, were computed using the software PowerMarker (Liu and Muse 2005). The proportion-of-shared-alleles distance (Dpsa; Bowcock et al., 1994) between pairs of accessions was calculated using the MICROSAT software (Minch et al., 1997). Cluster analysis based on distance matrix was performed using the Neighbor-Joining algorithm as implemented in NEIGHBOR program of the PHYLIP ver. 3.6b software package (Felsenstein, 2004). The reliability of the tree topology was assessed via bootstrapping (Felsenstein, 1985) over 1000 replicates generated by MICROSAT and subsequently used in NEIGHBOR and CONSENSE program in PHYLIP.

2.3.6. Cleaved Amplified Polymorphic Sequence and derived Cleaved Amplified Polymorphic Sequence markers for mapping

Using the sequence information of the cross-amplified monomorphic markers from both L. cicera RILs parental lines,

41 Chapter 2 ______restriction regions overlapping single nucleotide polymorphisms (SNPs) where detected using the Geneious software (Drummond et al., 2011) in order to design polymorphic cleaved amplified polymorphic sequence CAPS markers. When there were no restriction sites suitable to design CAPS markers for the SNP screening, derived cleaved amplified polymorphic sequence (dCAPS) markers were designed using the web based program dCAPS Finder 2.0 (Neff et al., 2002). See Table 2.2 for details on the SNPs, restriction enzymes, CAPS and dCAPS designed primers. PCR products were digested with the suitable enzyme (FastDigest, Fermentas), following the manufactures’ instructions. Digestion products were then resolved in 1.75 % Seaken LE agarose gel (Lonza, Rockland, USA) with SYBRSafe (Invitrogen, Eugene, USA) and visualised using a GEL-DOC 1000 System (Bio Rad, Hercules, USA).

2.4. Results and discussion

This study was performed to provide new molecular tools for L. cicera and L. sativus, since the few specific molecular markers developed so far for the construction of a linkage map in L. sativus (Skiba et al., 2003) were unsuitable to perform molecular studies in our Lathyrus spp. working accessions. One important constrain when using those molecular markers was that the 17 markers tested were all monomorphic between our mapping accessions of interest (BGE008277, BGE023542 and BGE015746, BGE024709) (data not shown). In order to obtain these new tools, we determined the transferability rate of different types of molecular markers developed for Lathyrus related species and tested their applicability for mapping and diversity studies in Lathyrus spp.

42 Table 2.2 – Characteristics of developed CAPS and dCAPS markers. ______

Marker Restriction Ta Marker name Forward primer (5'–3')a Reverse primer (5'–3')a type Enzyme (ºC)

Psmt_EST_00196_01_1 CAPS NlaIII TGGATTGGTAGGAGACTCGG TAGTTATGCTCATGGCGCTG 55 SHMT CAPS HinfI TTGCTGAGAACCTGCTCTTGGTATG ACCACAACTCACAAGTCACTTC 58 Pis_GEN_7 _1_2 _1 CAPS AluI AGTGCTCCTAGCATAGCCCA TTGATGGCAGCTACAGCAAG 60 Pis_GEN_7 _1_2 _1 CAPS DraI AGTGCTCCTAGCATAGCCCA TTGATGGCAGCTACAGCAAG 60 tRALS CAPS AluI GCAATTCCCTCCTCAGCTAAAAGTG GGTCTGCGAGCTGTTTTTGGAGAAG 60 tRALS CAPS RsaI GCAATTCCCTCCTCAGCTAAAAGTG GGTCTGCGAGCTGTTTTTGGAGAAG 60 mtmt_GEN_01947_02_1 CAPS MseI GGAGCACGGTGTTCTCTCTC GTAAAGTTGCCAGCCCATGT 60 mtmt_GEN_00036_02_1_dCAPS dCAPS HinfI CAAAAGGAATTGCACGTTACTG AATGACGGTAGCTTGGATGG 43 SHMT_dCAPS dCAPS AluI GCTAAACCGGTAATGGTATAG ACCACAACTCACAAGTCACTTC 45 Lup185_dCAPS dCAPS RsaI CCACTTTCCATGACAAATCT AAAAATTAGTATGGTCCAGTA 45 Peaβglu_dCAPS dCAPS AluI CTCTTTCCCTCCATCTAAAAG AAACAACAACATTCACCCAACC 45 Peaβglu_dCAPS dCAPS DraIII ACAGTATGTTCTACCTGCAACACAAAG AAACAACAACATTCACCCAACC 45 DRR206-d_dCAPS dCAPS HpaI CACCTCAATCCCACTTTGT AAAGAACTTGATATAAACACC 47

abase mismatch introduced in dCAPS primers are in bold and underlined Cross - amplification species 43

Chapter 2 ______

2.4.1. Transferability of microsatellites to Lathyrus spp.

EST-SSR markers (in this case only developed for pea) were the ones presenting the higher rates of simple pattern amplification on the two Lathyrus spp. (33.3 % and 42.9 % on L. cicera and L. sativus respectively) from all the tested marker types (Figure 2.1). These values dropped quite dramatically when analysing the general results of the gSSR markers (Figure 2.1). From the lentil gSSR tested, not even one marker successfully amplified one single fragment in L. sativus.

The two Lathyrus species presented very similar results when analysing the non-simple patterns of amplification obtained with the pea EST-SSR and the faba bean gSSR (Figure 2.1). Both species had in general higher percentages of SSRs resulting in complex amplification pattern then failed to amplify any fragment. These differences were not so obvious in the case of the pea gSSR. In the case of the lentil gSSRs, both Lathyrus sp. had higher percentages of failed amplification (Figure 2.1).

Figure 2.1 - Numbers and percentages of gSSRs and EST-SSRs from lentil, pea and faba bean amplified in Lathyrus spp..

44 Cross-species amplification ______

From all the markers tested, only the pea EST-SSR PSBLOX13.2 was polymorphic between the L. sativus parental accessions (BGE015746, BGE024709), using a 4 % Metaphor gel.

The observed transferability rates across different genera of the cross-amplified microsatellites were in accordance with previous studies for EST-SSRs and gSSRs (Peakall et al., 1998; Gupta et al., 2003; Gutierrez et al., 2005; Varshney et al., 2005; Perez et al., 2006). Those authors reported transferability values that varied from 18-78 % for EST-SSRs and 3-24 % for gSSRs. The higher transferability rate for EST-SSRs was expected due to their location in coding regions. This feature makes them more transferable than gSSRs, but also less polymorphic among individuals of the same species (Gupta et al., 2003). Since both L. sativus and L. cicera are diploid, EST-SSRs and gSSRs showing complex band patterns with more than two alleles implied that these loci arise from duplication events. In the case of EST-SSRs this may suggest the existence of multigene families (Raji et al., 2009). In the case of the gSSRs, their location in the Long Terminal Repeat (LTR) of ancient retrotransposons could be the reason for the complex band patterns encountered, since these regions are prone to duplications (Smýkal et al., 2009).

To confirm the specificity of the amplified regions and the presence of the microsatellite motifs on the amplicons, we aligned the two L. cicera parental accessions (BGE008277 and BGE023542) and one L. sativus (BGE015746) sequenced fragments against the reference sequence from the donor species using the Geneious

45 ______Chapter 2 46 . Table 2.3 – Characteristics of pea microsatellite markers successfully amplified and sequenced on Lathyrus spp. GenBank information type: (g) genomic sequence; (m) mRNA sequence

% pairwise identity Top BLASTn hit Lathyrus Marker Target P. sativum Top Primer name amplicon Pea SSR motif Lathyrus SSR motif Pea GenBank type species sequenced BLASTn E-value Size (bp) Accession amplicon hit Lathyrus 3,31E- AA241 gSSR 283 (TC)4 (TC)2 - 75,4 CU655881 (g) cicera 40 Lathyrus 4,92E- AA241 gSSR 284 (TC)4 (TC)2 - 74,7 CU655881 (g) sativus 38 Lathyrus AB111 gSSR 139 (AG)15 (AG)3 72,2 - - - sativus Lathyrus (ATCTCT)7ATCTAT AD160 gSSR 136 TAG(CT)4(AT)4 68,1 - - - cicera (CT)5(ATCT)9 Lathyrus (ATCTCT)7ATCTAT AD160 gSSR 132 TAG(CT)5(AT)4 57,9 - - - sativus (CT)5(ATCT)9 EST- Lathyrus 1,58E- AA427337 193 (AC)5 (AC)4 - 85,2 AA427337 (m) SSR cicera 53 EST- Lathyrus 2,57E- AA427337 192 (AC)5 (AC)4 - 89,4 AA427337 (m) SSR sativus 63 EST- Lathyrus 3,60E- AA430902 188 (AAT)7(AT)3 (AAT)5(TA)2 - 89,1 AA430942 (m) SSR cicera 56 EST- Lathyrus 1,05E- AA430902 185 (AAT)7(AT)3 (TAA)3(TA)3 - 86,3 AA430942 (m) SSR sativus 61 EST- Lathyrus 3,60E- AA430942 188 (AAT)7(AT)3 (AAT)5(TA)2 - 89,1 AA430942 (m) SSR cicera 56 EST- Lathyrus 1,05E- AA430942 185 (AAT)7(AT)3 (TAA)3(TA)3 - 86,3 AA430943 (m) SSR sativus 61 EST- Lathyrus 4,61E- CHPSGPA1 193 (AT)17(AGATAT)2 (AT)6AGATAT - 84,8 X15190 (m) SSR cicera 67 EST- Lathyrus 9,91E- CHPSGPA1 189 (AT)17(AGATAT)2 (AT)4AGATAT - 83,0 X15190 (m) SSR sativus 63 EST- Lathyrus 1,16E- CHPSTZPP 373 (TA)6 (TA)4 - 70,2 X56315 (g) SSR cicera 40

______Table 2.3 (cont.) % pairwise identity Lathyrus Marker Target Pea GenBank Primer name amplicon Pea SSR motif Lathyrus SSR motif P. sativum Top E-value type species Accession Size (bp) sequenced BLASTn amplicon hit EST- Lathyrus 2,40E- ______CHPSTZPP 303 (TA)6 (TA)4 - 84,3 X56315 (g) SSR sativus 30 EST- Lathyrus 1,05E- PEAATPSYND 209 (AC)5 (AC)4 - 91,0 M94558 (m) SSR cicera 75 EST- Lathyrus 1,28E- PEAATPSYND 209 (AC)5 (AC)4 - 90,5 M94558 (m) SSR sativus 74 EST- Lathyrus 1,59E- PEACPLHPPS 114 (TA)3(AT)6 (AT)2 - 84,1 L19651 (m) SSR cicera 31 EST- Lathyrus 1,91E- PEACPLHPPS 113 (TA)3(AT)6 (AT)2 - 84,1 L19651 (m) SSR sativus 30 EST- Lathyrus 1,11E- PEAOM14A 197 (CCT)6 CCT…(CCT)3 - 90,9 M69105 (m) SSR cicera 68 EST- Lathyrus 4,90E- PEAOM14A 203 (CCT)6 CCTGCTCCT…(CCT)3 - 88,2 M69105 (m) SSR sativus 67 EST- Lathyrus 6,35E- PSAJ3318 172 (TA)3…(CAT)6 (TA)9…(CAT)4 - 86,0 AJ223318 (m) SSR cicera 52 EST- Lathyrus 2,74E- PSAJ3318 165 (TA)3…(CAT)6 (TA)7…(CAT)4 - 88,5 AJ223318 (m)

SSR sativus 56 Cross EST- Lathyrus 3,44E- PSAS 235 (AAT)6 (AAT)4 - 86,2 Y13321 (g) SSR sativus 76 -

EST- Lathyrus 2,09E- amplification species PSBLOX13.2 92 (CAT)8 (CAT)4 - 88,3 X78581 (g) SSR cicera 28 EST- Lathyrus 9,04E- PSBLOX13.2 93 (CAT)8 (CAT)5 - 87,4 X78581 (g) SSR sativus 27 EST- Lathyrus CCTCTT(CCT)3 2,20E- PSBT2AGEN 266 (CCT)5(CTT)2CAT(CTT)3 - 95,5 X96764 (m) SSR cicera (CTT)2CAT(CTT)2 117 EST- Lathyrus CCTCTT(CCT)3 3,81E- PSBT2AGEN 266 (CCT)5(CTT)2CAT(CTT)3 - 96,7 X96764 (m) SSR sativus (CTT)2CAT(CTT)2 121 4 7

______Chapter 2 48

Table 2.3 (cont.) % pairwise identity Lathyrus Marker Target Pea GenBank Primer name amplicon Pea SSR motif Lathyrus SSR motif P. sativum Top E-value type species Accession Size (bp) sequenced BLASTn amplicon hit EST- Lathyrus 2,07E- PSGAPA1 147 (AT)17(AGATAT)2 (AT)6AGATAT - 81,5 X15190 (m) SSR cicera 44 EST- Lathyrus 1,54E- PSGAPA1 142 (AT)17(AGATAT)2 (AT)4AGATAT - 79,1 X15190 (m) SSR sativus 39 EST- Lathyrus 2,12E- PSGDPP 206 AACATC(AAC)5 (AAC)2TAT(AAC)2 - 94,6 X59773 (m) SSR cicera 84 EST- Lathyrus 2,57E- PSGDPP 205 AACATC(AAC)5 (AAC)2TAT(AAC)2 - 94,1 X59773 (m) SSR sativus 83 EST- Lathyrus 2,04E- PSJ000640A 199 (AAC)7 (AAC)5 - 82,0 AJ000640 (g) SSR cicera 46 EST- Lathyrus 3,50E- PSJ000640A 194 (AAC)7 (AAC)5 - 80,2 AJ000640 (g) SSR sativus 42 EST- Lathyrus 1,07E- PSU51918 136 (GAA)6 (GAA)4 - 91,6 U51918 (m) SSR cicera 48 EST- Lathyrus 1,59E- PSU51918 134 (GAA)6 (GAA)3 - 89,5 U51918 (m) SSR sativus 46 EST- Lathyrus (TGG)2 CGG(TGG)2 3,75E- PSU81287 265 (TGG)5TTAT(TGG)3 - 96,0 U81287 (m) SSR cicera TTAT(TGG)3 37 EST- Lathyrus 3,52E- PSU81287 269 (TGG)5TTAT(TGG)3 (TGG)5TTAT(TGG)3 - 94,6 U81287 (m) SSR sativus 31 EST- Lathyrus 1,90E- PSZINCFIN 210 (AAC)5 (AAC)5 - 96,2 X87374 (m) SSR cicera 91 EST- Lathyrus 1,90E- PSZINCFIN 210 (AAC)5 (AAC)5 - 96,2 X87374 (m) SSR sativus 91

Cross-species amplification ______alignment software (Drummond et al., 2011). For the donor fragments not found in the exploited databases, amplicons from the donor species were also sequenced to allow comparison (Table 2.3).

Only one cross-amplified pea gSSR loci (AA241), sequenced in the Lathyrus spp., gave rise to a significant BLAST hit, with an E- value of 3.3e-40 for L. cicera and 4.9e-38 for L. sativus, allowing comparison of the microsatellite motif with pea. Five other cross- amplified gSSR fragments lacking a BLAST hit had also to be sequenced in pea to allow comparisons. In two of these amplicons (AB111, AD160) the sequences were similar to the donor pea despite the low pairwise identity (58 – 72 %), because although the SSR flanking regions were conserved, a large portion of the SSR region was missing in the two Lathyrus spp. (Table 2.3). In the other three sequenced loci (AD146, D21 and SSR124) in Lathyrus spp., the amplified regions were not conserved. The sequenced amplicons of the pea EST-SSRs, producing one single fragment in L. cicera and L. sativus, showed that the cross-amplified loci were orthologous to the donor species (maximum E-value = 9e-27). Microsatellite size homoplasy, confirmed by sequence alignment, was present in two markers (PEAATPSYND, PSZINCFIN) when comparing pea with both Lathyrus spp.. Additionally, PSGDPP had the same sequence length in pea and L. cicera. In the case of PSBT2AGEN microsatellite size homoplasy was only detected between both Lathyrus spp.. For other two pea EST-SSRs the repeat motif was maintained (PSU81287 in L. sativus and PSZINCFIN in both Lathyrus spp.). In most of the cases there were more than one mutation event in each microsatellite region (Table 2.3). When comparing the mutations in those loci, only in PSGDPP the mutations were equal for both

49 Chapter 2 ______species. The most common mutations in the microsatellite motifs were deletions (57.1 %) (Table 2.3).

2.4.2. Transferability of Intron-Targeted Amplified Polymorphic markers, Defence-Related and Resistance Gene Analogues to Lathyrus spp.

Differences in the transferability rates among ITAP, DR and RGA markers were much less pronounced than between EST-SSR and gSSR. Both Lathyrus spp. had very similar results in the transferability rates of each of these particular molecular marker types (Figure 2.2). The majority of the tested markers resulted in one single fragment amplification with the highest rates for the ITAP markers (58% in both Lathyrus sp.), follow by the pea DR (58%) and finally the pea RGA (50%) (Figure 2.2).

Few of the tested markers failed to amplify in both Lathyrus sp., but 25 % to 40 % of each marker types resulted in a complex pattern of amplification (Figure 2.2). Thirteen ITAPs presented a direct polymorphism between two of the parental accessions tested, four only in L. cicera and eight only in L. sativus (ESM 2.1). Additionally, one ITAP (Lup280) was polymorphic between both species’ parental accessions.

50 Cross-species amplification ______

Figure 2.2 Numbers and percentages of ITAPs and pea DRs and RGAs amplified in Lathyrus spp..

From the 58 ITAP markers giving a single amplicon in L. cicera that were sequenced, 28 presented high homology (>70 % identity) with the donor sequence (ESM 2.2). In addition, 10 ITAP amplicons had a BLASTn hit with low pairwise identity (below 70%). The reason for this was that the sequence present in the NCBI database is an mRNA molecule that, when aligned with our genomic sequence, misses the intronic region. Nevertheless, five other ITAP amplicons, presenting large insertions or deletions in one or two sections, were considered as positive hits, since the aligned exonic regions were highly conserved. All these L. cicera amplicons were BLASTed against the NCBI database, where 55 (94.8 %) presented an E-value < 1e-10 (ESM 2.2).

From the 61 ITAP marker fragments sequenced in L. sativus (ESM 2.2), 43 where considered homologous to the donor species, 25 of which presenting a high pairwise identity (>70 %) (comparison between DNA and RNA). All the amplicons were BLASTed against

51 Chapter 2 ______the NCBI database, where 95.1 % presented an E-value < 1e-10 (ESM 2.2).

Not all amplified sequences from the DR and RGA could be compared to the reference donor species due to the lack of information about the donor’ sequence in databases and the failure to be sequenced after several attempts. The ones that could be compared (three DRs and two RGAs) showed a pairwise identity above 70 % (ESM 2.2). Nevertheless all the cross-amplified DR and RGA amplicons had a significant BLAST hit (maximum E-value = 1.34e-38). From the DR and RGA markers originating one single amplicon (Figure 2.2), just the DR DRR230-d presented a direct polymorphism between the L. sativus parental accessions.

2.4.3. Usefulness of cross amplified markers on Lathyrus spp. linkage mapping

In order to access the usefulness for linkage mapping, the successfully cross-amplified markers, were analysed for size polymorphism and SNPs among the Lathyrus parental accessions. Segregation ratios of polymorphic markers were then tested in a L. cicera RILs population segregating for rust and powdery mildew.

From the transferable markers, one EST-SSR, one DR and nine ITAPs presented direct size polymorphism between the parental accessions of the L. sativus mapping population (ESM 2.1). Additionally five ITAP markers were size polymorphic between the L. cicera parental accessions (ESM 2.1) and were used to screen the L. cicera RILs mapping population. In this screening all the molecular marker segregations presented a χ2 value between 0.08 and 1.52

52 Cross-species amplification ______

(α=0.05) confirming that the segregation ratios were not deviating from the expected 1:1 segregation (Table 2.4).

For the markers presenting microsatellite size homoplasy in L. cicera (ESM 2.2), 29 fragments amplified in both parental accessions were sequenced and aligned to detect SNPs and design CAPS. When no restriction sites were associated with SNPs, dCAPS markers were designed to allow their scoring in the L. cicera RILs population. The high sequence similarity existing among the accessions of L. cicera decreased the probability of detecting SNPs. Potential sequencing errors or a low GC content in the amplicon region, further hampered the design of suitable primer pairs. As result, just 10 fragments containing SNPs were detected and for only nine of these it was possible to design CAPs or dCAPS markers. From the seven CAPs identified in five fragments and six dCAPS developed for other five fragments (Table 2.2) only three and one markers respectively, were suitable to be screened in the RILs L. cicera population and presented a 1:1 Mendelian segregation (Table 2.4).

The other four CAPs failed to

53 Chapter 2 ______give clear polymorphisms in the parental accessions and the other five dCAPS did no amplified at all or there was no cleavage of the amplicon (Table 2.3). All of these unsuitable markers were not screened in the RILs.

The use of alternative SNP screening approaches (such as dHPLC, ecoTILLING or TaqMan) not depending on specific sequences for restriction endonucleases would have allowed the identification of a higher number of SNPs to genotype in the L. cicera segregation population.

The inclusion of codominant molecular markers, as the ones here identified, on the Lathyrus spp. linkage maps will allow the future identification of chromosomal rearrangements with different donor species. The utility of transferable markers to access macrosynteny (Phan et al., 2007; Ellwood et al., 2008; Hougaard et al., 2008) and microsynteny (Gualtieri et al., 2002; Guyot et al., 2012) in related and distant species have been already shown by several previous studies.

2.4.4. Usefulness of cross amplified markers on Lathyrus spp. diversity studies

In order to assess the utility of the cross amplified markers for diversity studies a set of selected gSSRs, EST-SSRs and ITAPs were tested (Table 2.5). One of the selected gSSRs markers amplified in the Lathyrus sp. revealed a sequence very similar to the donor pea species although a large part of the microsatellite region was missing. As a result, the pea amplicon was almost double than that of the Lathyrus accessions and displayed a low pairwise identity value (AD160 – 58 %). A second selected gSSR marker had a similar size and a higher pairwise identity value when comparing to the donor pea

54 Cross-species amplification ______

(AA241 – 73 %). The selected ITAPs and EST-SSRs displayed a similar size when compared with the donor species. In total, 2 EST- SSRs, 2 gSSRs and 13 ITAPs, were tested in 20 random Lathyrus spp. individuals, 10 L. cicera and 10 L. sativus.

For four of the seven monomorphic markers in both species (Table 2.5), L. cicera presented a different allele to L. sativus (AA241, AD160, LG054 and mt_00495_01_1). Using the selected markers, the higher polymorphic information content (PIC) was observed for the L. sativus accessions (mean = 0.174). Two markers were highly informative for L. sativus (Pis_GEN_21_1_1 and PSBLOX13.2) with a PIC > 0.6. For the L. cicera accessions under study there was no heterozygosity detected which explains the lower PIC values obtained (mean = 0.110) (Table 2.5).

The diversity analysis of the cross-amplified molecular markers presented here, clearly distinguishes between L. cicera and L. sativus individuals (Figure 2.3). Among L. cicera accessions, the genetic distance (DPSA) varied from 0.000 to 0.435, with a mean value of 0.167. Among L. sativus accessions, the distance varied from 0.065 to 0.375 with a mean value of 0.203. A total of four (BGE001043, BGE001164, BGE023542, BGE023558) and two (BGE022223, BGE027064) indistinguishable accessions were observed, all from L. cicera. The mean genetic distance between species was 0.817 ranging from 0.782 to 0.877. Also, the mapping parental lines from the L. cicera RIL population (BGE008277 and BGE023542) were well separated in the Neighbor-Joining tree.

55 56 ______Chapter Table 2.5 – Lathyrus cicera and Lathyrus sativus cross amplifiable markers diversity analysis

Marker Lathyrus cicera Lathyrus sativus Marker

a b c d e f a b c d e f 2 type N Na Allele size MAF HE HO PIC N Na Allele size MAF HE HO PIC

LG041 ITAP MLG 10 1 331 - - - 0,000 10 1 331 - - - 0,000

LG054 ITAP MLG 10 1 281 - - - 0,000 10 1 278 - - - 0,000

Lup280 ITAP ML 10 3 578, 602, 622 0,100 0,540 0,000 0,466 10 3 580, 588, 632 0,050 0,515 0,300 0,424

mtmt_GEN_00012_03_1 ITAP GLIP 10 2 658, 670 0,200 0,320 0,000 0,269 10 2 232, 235 0,450 0,495 0,900 0,372

mtmt_GEN_00385_06_1 ITAP GLIP 10 2 604, 644 0,300 0,420 0,000 0,332 10 2 629, 633 0,050 0,095 0,100 0,090

mtmt_GEN_00423_02_2 ITAP GLIP 10 1 355 - - - 0,000 10 2 389, 391 0,400 0,480 0,400 0,365

mtmt_GEN_00495_01_1 ITAP GLIP 10 1 475 - - - 0,000 10 1 463 - - - 0,000

mtmt_GEN_01115_02_1 ITAP GLIP 10 2 329, 337 0,200 0,320 0,000 0,269 10 1 337 - - - 0,000

Pis_GEN_21_1_1 ITAP GLIP 10 1 597 - - - 0,000 10 6 534, 537, 543, 549, 576, 579 0,050 0,725 0,700 0,692

Pis_GEN_6_3_1 ITAP GLIP 10 1 500 - - - 0,000 10 2 602, 638 0,400 0,480 0,200 0,365

Pis_GEN_9_3_1 ITAP GLIP 10 1 500 - - - 0,000 10 1 500 - - - 0,000

Pis_PR_127_1 ITAP GLIP 10 2 419, 422 0,200 0,320 0,000 0,269 ------

Psat_EST_00178_01_1 ITAP GLIP 10 2 362, 410 0,200 0,320 0,000 0,269 10 1 410 - - - 0,000

PSBLOX13.2 EST-SSR 10 1 90 - - - 0,000 10 5 93, 96, 99, 102, 105 0,050 0,695 0,900 0,643

PSGDPP EST-SSR 10 1 205 - - - 0,000 10 1 205 - - - 0,000

AA241 gSSR 10 1 283 - - - 0,000 10 1 284 - - - 0,000

AD160 gSSR 10 1 135 - - - 0,000 10 1 133 - - - 0,000

Mean - 1,4 - 0,200 0,373 0,000 0,110 - 1,9 - 0,207 0,498 0,500 0,185

anumber of analysed individuals; bnumber of alleles; cminor allele frequency; dexpected heterozygosity; eobserved heterozygosity; fpolymorphism information content Cross-species amplification ______

Previous studies on Lathyrus diversity employed a range of different molecular markers like Random Amplified Polymorphic DNAs (RAPD) (Croft et al., 1999), Isozymes (Chowdhury and Slinkard, 2000; Ben Brahim et al., 2002; Gutiérrez-Marcos et al., 2006), Inter-Simple Sequence Repeats (ISSR) (Belaid et al., 2006), Amplified Fragment Length Polymorphisms (AFLP) (Tavoletti and Iommarini, 2007; Lioi et al., 2011) and also EST-SSRs (Lioi et al., 2011; Shiferaw et al., 2011). In contrast with the only work analysing L. sativus accessions (Croft et al., 1999), the present study, using the set of selected cross-amplified molecular markers, managed to distinguish all L. sativus accessions. Nevertheless this comparison should be made with caution since more individuals from each accession and a different molecular marker type (RAPD), known for their higher polymorphism rate, were used on the previous study. When comparing with the isozyme analysis reported by Gutierrez- Marcos (2006), similar Nei’s genetic identity average values (Nei, 1973) were obtained between the accessions from Iberian origin (0.888) and the ones considered in this work (0.864 for L. cicera and 0.886 for L. sativus), thus confirming the existence of high genetic diversity and a good representation of the Lathyrus Iberian germplasm diversity in our 20 accessions.

To our knowledge, only two studies using cross-amplified EST-SSRs in Lathyrus have been reported so far. One of them used EST-SSRs developed for M. truncatula to access the diversity in Ethiopian L. sativus populations (Shiferaw et al., 2011) and the second used EST-SSRs developed for L. japonicus to discriminate between different L. sativus accessions from Italian origin (Lioi et al., 2011).

57 Chapter 2 ______

Figure 2.3 - Neighbor-Joining tree based on the proportion-of-shared-alleles distance values among 20 Lathyrus individuals. Numbers at nodes indicate bootstrap values (%) out of 1.000 replications; only values above 50 % are shown. Black dots indicate the L. cicera RILs parental lines.

Our work expand upon the previous works in two ways; first, by testing the cross-amplification of a much higher number of different types of markers from different related legume species and, second, all of the cross-amplified SSRs and most of the ITAPs, DRs and RGAs were confirmed by sequencing in order to verify the correspondence between the donor and the target species sequences.

58 Cross-species amplification ______

2.5. Conclusions

This paper describes the transferability of molecular markers from Medicago truncatula, Pisum sativum, Lens culinaris, Lupinus spp. and Vicia faba to Lathyrus spp. for their application in mapping and diversity studies. Cross-genera amplification of molecular markers provides an alternative for the development of new molecular markers on understudied genus. The increase availability of public legume resources constitutes an efficient and cost-effective source of molecular markers for use in Lathyrus breeding and genetics. Nevertheless, one of the primary factors limiting broader cross- markers applications is the unsuccessful amplification of homologous products across species. It is therefore, advisable to know the sequence of the amplified fragments if the objective is to draw conclusions about equivalent genomic regions between species.

Our results revealed quite high marker transferability among the above mentioned legume species. The number of transferred markers would likely improve by using mainly markers developed from expressed sequences. EST-derived SSRs, which can easily be extracted from EST databases, can be successfully used for this purpose. They will also have a higher probability of being in linkage disequilibrium with genes/QTLs controlling economic traits, thus proving relatively more useful for studies involving marker-trait association, QTL mapping and genetic diversity analysis.

Additionally in the future, due to the constant diminishing sequencing costs, specific Lathyrus cicera or L. sativus genomic libraries will also provide gene based molecular markers (EST-SSRs and SNP markers) for these species. These new markers could be annotated using the information already available for related species such as Medicago truncatula or Glycine max, or mapped on other

59 Chapter 2 ______major legume crops linkage maps not yet totally sequenced as pea, faba beans and chickpea, as a natural complementation of the now generated marker information.

Our outcomes have increased the number of molecular markers available for Lathyrus species, and particularly for our L. cicera and L. sativus crosses which did not had yet useful codominant markers for developing a linkage map useful in future genetics studies. A Lathyrus linkage map containing these cross-amplified markers will establish the basis to enable future comparative mapping across legume species.

2.6. Acknowledgements

NFA carried out the cross-amplifications, sample processing for sequencing, data analysis and drafted the manuscript. STL participated in the results generation on genotyping the L. cicera RILs and the Lathyrus spp. accessions for the diversity study. DR generated the RIL populations studied. CC, AMT and DR revised the manuscript critically. MCVP coordinated the study and participated to the drafting and revision of the manuscript. NFA would like to thank CRF-INIA, Madrid, Spain, for supplying the accessions, Mara Lisa Alves for helping in the diversity study analysis and Alberto Martín- Sanz for the technical support on the primer pairs supply.

2.7. References

Asmussen, C.B., and Liston, A. (1998). Chloroplast DNA characters, phylogeny, and classification of Lathyrus (Fabaceae). American Journal of Botany 85, 387-401. Belaid, Y., Chtourou-Ghorbel, N., Marrakchi, M., and Trifi-Farah, N. (2006). Genetic diversity within and between populations of

60 Cross-species amplification ______

Lathyrus genus (Fabaceae) revealed by ISSR markers. Genetic Resources and Crop Evolution 53, 1413-1418. Ben Brahim, N., Combes, D., and Marrakchi, M. (2001). Autogamy and allogamy in genus Lathyrus. Lathyrus Lathyrism Newsletter 2, 21-26. Ben Brahim, N., Salhi, A., Chtourou, N., Combes, D., and Marrakchi, M. (2002). Isozymic polymorphism and phylogeny of 10 Lathyrus species. Genetic Resources and Crop Evolution 49, 427-436. Bennett, M.D., and Leitch, I.J. (2012). "Angiosperm DNA C-values database". (release 8.0, Dec. 2012) http://data.kew.org/cvalues/). Bowcock, A.M., Ruiz-Linares, A., Tomfohrde, J., Minch, E., Kidd, J.R., and Cavalli-Sforza, L.L. (1994). High resolution of human evolutionary trees with polymorphic microsatellites. Nature 368, 455-457. Burstin, J., Deniot, G., Potier, J., Weinachter, C., Aubert, G., and Baranger, A. (2001). Microsatellite polymorphism in Pisum sativum. Plant Breeding 120, 311-317. Calderón, F.J., Vigil, M.F., Nielsen, D.C., Benjamin, J.G., and Poss, D.J. (2012). Water use and yields of no-till managed dryland grasspea and yellow pea under different planting configurations. Field Crops Research 125, 179-185. Campbell, C.G. (1997). Promoting the conservation and use of underutilized and neglected crops. Grass pea. Lathyrus sativus L. Rome, Italy: International Plant Genetic Resources Institute. Castillo, A., Budak, H., Varshney, R.K., Dorado, G., Graner, A., and Hernandez, P. (2008). Transferability and polymorphism of barley EST-SSR markers used for phylogenetic analysis in Hordeum chilense. BMC Plant Biology 8, 97. Choi, H.K., Mun, J.H., Kim, D.J., Zhu, H.Y., Baek, J.M., Mudge, J., Roe, B., Ellis, N., Doyle, J., Kiss, G.B., Young, N.D., and Cook, D.R. (2004). Estimating genome conservation between crop and model legume species. Proceedings of the National Academy of Sciences of the United States of America 101, 15289-15294. Chowdhury, M.A., and Slinkard, A.E. (1999). Linkage of random amplified polymorphic DNA, isozyme and morphological markers in grasspea (Lathyrus sativus). Journal of Agricultural Science 133, 389-395. Chowdhury, M.A., and Slinkard, A.E. (2000). Genetic diversity in grasspea (Lathyrus sativus L.). Genetic Resources and Crop Evolution 47, 163-169.

61 Chapter 2 ______

Croft, A.M., Pang, E.C.K., and Taylor, P.W.J. (1999). Molecular analysis of Lathyrus sativus L. (grasspea) and related Lathyrus species. Euphytica 107, 167-176. Datta, S., Mahfooz, S., Singh, P., Choudhary, A., Singh, F., and Kumar, S. (2010). Cross-genera amplification of informative microsatellite markers from common bean and lentil for the assessment of genetic diversity in pigeonpea. Physiology and Molecular Biology of Plants 16, 123-134. Drummond, A.J., Ashton, B., Buxton, S., Cheung, M., Cooper, A., Duran, C., Field, M., Heled, J., Kearse, M., Markowitz, S., Moir, R., Stones-Havas, S., Sturrock, S., Thierer, T., and Wilson, A. (2011). "Geneious v5.4", Available from http://www.geneious.com/). Ellwood, S.R., Phan, H.T.T., Jordan, M., Hane, J., Torres, A.M., Avila, C.M., Cruz-Izquierdo, S., and Oliver, R.P. (2008). Construction of a comparative genetic map in faba bean (Vicia faba L.); conservation of genome structure with Lens culinaris. BMC Genomics 9, 380. Felsenstein, J. (1985). Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783-791. Felsenstein, J. (2004). "PHYLIP (Phylogeny Inference Package) version 3.6. Distributed by the author. , Seattle, WA Department of Genomic Sciences, University of Washington".). Feng, S.P., Li, W.G., Huang, H.S., Wang, J.Y., and Wu, Y.T. (2009). Development, characterization and cross-species/genera transferability of EST-SSR markers for rubber tree (Hevea brasiliensis). Molecular Breeding 23, 85-97. Fondevilla, S., Almeida, N.F., Satovic, Z., Rubiales, D., Vaz Patto, M.C., Cubero, J., and Torres, A.M. (2011). Identification of common genomic regions controlling resistance to Mycosphaerella pinodes earliness and architectural traits in different pea genetic backgrounds. Euphytica 182, 43-52. Fondevilla, S., Fernández-Aparicio, M., Satovic, Z., Emeran, A.A., Torres, A.M., Moreno, M.T., and Rubiales, D. (2010). Identification of quantitative trait loci for specific mechanisms of resistance to Orobanche crenata Forsk. in pea (Pisum sativum L.). Molecular Breeding 25, 259-272. Fondevilla, S., Satovic, Z., Rubiales, D., Moreno, M.T., and Torres, A.M. (2008). Mapping of quantitative trait loci for resistance to Mycosphaerella pinodes in Pisum sativum subsp syriacum. Molecular Breeding 21, 439-454. Grela, E., Rybiński, W., Matras, J., and Sobolewska, S. (2012). Variability of phenotypic and morphological characteristics of

62 Cross-species amplification ______

some Lathyrus sativus L. and Lathyrus cicera L. accessions and nutritional traits of their seeds. Genetic Resources and Crop Evolution 59, 1687-1703. Gualtieri, G., Kulikova, O., Limpens, E., Kim, D.J., Cook, D.R., Bisseling, T., and Geurts, R. (2002). Microsynteny between pea and Medicago truncatula in the SYM2 region. Plant Molecular Biology 50, 225-235. Gupta, P.K., Rustgi, S., Sharma, S., Singh, R., Kumar, N., and Balyan, H.S. (2003). Transferable EST-SSR markers for the study of polymorphism and genetic diversity in bread wheat. Molecular Genetics and Genomics 270, 315-323. Gusmao, M., Siddique, K.H.M., Flower, K., Nesbitt, H., and Veneklaas, E.J. (2012). Water deficit during the reproductive period of grass pea (Lathyrus sativus L.) reduced grain yield but maintained seed size. Journal of Agronomy and Crop Science 198, 430-441. Gutiérrez-Marcos, J.F., Vaquero, F., Sáenz De Miera, L.E., and Vences, F.J. (2006). High genetic diversity in a world-wide collection of Lathyrus sativus L. revealed by isozymatic analysis. Plant Genetic Resources 4, 159-171. Gutierrez, M.V., Vaz Patto, M.C., Huguet, T., Cubero, J.I., Moreno, M.T., and Torres, A.M. (2005). Cross-species amplification of Medicago truncatula microsatellites across three major pulse crops. Theoretical and Applied Genetics 110, 1210-1217. Guyot, R., Lefebvre-Pautigny, F., Tranchant-Dubreuil, C., Rigoreau, M., Hamon, P., Leroy, T., Hamon, S., Poncet, V., Crouzillat, D., and De Kochko, A. (2012). Ancestral synteny shared between distantly-related plant species from the asterid (Coffea canephora and Solanum sp.) and rosid (Vitis vinifera) clades. BMC Genomics 13, 103. Hamwieh, A., Udupa, S.M., Choumane, W., Sarker, A., Dreyer, F., Jung, C., and Baum, M. (2005). A genetic linkage map of Lens sp. based on microsatellite and AFLP markers and the localization of fusarium vascular wilt resistance. Theoretical and Applied Genetics 110, 669-677. Hanbury, C.D., White, C.L., Mullan, B.P., and Siddique, K.H.M. (2000). A review of the potential of Lathyrus sativus L. and L. cicera L. grain for use as animal feed. Animal Feed Science and Technology 87, 1-27. Harris-Shultz, K., Milla-Lewis, S., and Brady, J. (2012). Transferability of SSR and RGA Markers Developed in Cynodon spp. to Zoysia spp. Plant Molecular Biology Reporter 30, 1264-1269. Hougaard, B.K., Madsen, L.H., Sandal, N., Moretzsohn, M.D., Fredslund, J., Schauser, L., Nielsen, A.M., Rohde, T., Sato,

63 Chapter 2 ______

S., Tabata, S., Bertioli, D.J., and Stougaard, J. (2008). Legume anchor markers link syntenic regions between Phaseolus vulgaris, Lotus japonicus, Medicago truncatula and Arachis. Genetics 179, 2299-2312. Kumar, J., Pratap, A., Solanki, R.K., Gupta, D.S., Goyal, A., Chaturvedi, S.K., Nadarajan, N., and Kumar, S. (2012). Genomic resources for improving food legume crops. Journal of Agricultural Science 150, 289-318. Lioi, L., Sparvoli, F., Sonnante, G., Laghetti, G., Lupo, F., and Zaccardelli, M. (2011). Characterization of Italian grasspea (Lathyrus sativus L.) germplasm using agronomic traits, biochemical and molecular markers. Genetic Resources and Crop Evolution 58, 425-437. Loridon, K., Mcphee, K., Morin, J., Dubreuil, P., Pilet-Nayel, M.L., Aubert, G., Rameau, C., Baranger, A., Coyne, C., Lejeune- Hènaut, I., and Burstin, J. (2005). Microsatellite marker polymorphism and mapping in pea (Pisum sativum L.). Theoretical and Applied Genetics 111, 1022-1031. Martín-Sanz, A., Palomo, J., Pérez De La Vega, M., and Caminero, C. (2012). Characterization of Pseudomonas syringae pv. syringae isolates associated with bacterial blight in Lathyrus spp. and sources of resistance. European Journal of Plant Pathology 134, 205-216. Minch, E., Ruiz Linares, A., Goldstein, D., Feldman, M., and Cavalli- Sforza, L. (1997). MICROSAT, Version 1.5 b. Standford CA: Standford University. Neff, M.M., Turk, E., and Kalishman, M. (2002). Web-based primer design for single nucleotide polymorphism analysis. Trends in Genetics 18, 613-615. Nei, M. (1973). Analysis of gene diversity in subdivided populations. Proceedings of the National Academy of Sciences of the United States of America 70, 3321-3323. Nelson, M.N., Moolhuijzen, P.M., Boersma, J.G., Chudy, M., Lesniewska, K., Bellgard, M., Oliver, R.P., Swiecicki, W., Wolko, B., Cowling, W.A., and Ellwood, S.R. (2010). Aligning a new reference genetic map of Lupinus angustifolius with the genome sequence of the model legume, Lotus japonicus. DNA Research 17, 73-83. Parra-Gonzalez, L., Aravena-Abarzua, G., Navarro-Navarro, C., Udall, J., Maugham, P., Peterson, L., Salgo-Garrido, H., and Maureira Butler, I. (2012). Yellow lupin (Lupinus luteus L.) transcriptome sequencing: molecular marker development and comparative studies. BMC Genomics 13, 425.

64 Cross-species amplification ______

Peakall, R., Gilmore, S., Keys, W., Morgante, M., and Rafalski, A. (1998). Cross-species amplification of soybean (Glycine max) simple sequence repeats (SSRs) within the genus and other legume genera: Implications for the transferability of SSRs in plants. Molecular Biology and Evolution 15, 1275-1287. Perez, J.O., Dambier, D., Ollitrault, P., D'eeckenbrugge, G.C., Brottier, P., Froelicher, Y., and Risterucci, A.M. (2006). Microsatellite markers in Carica papaya L.: isolation, characterization and transferability to Vasconcellea species. Molecular Ecology Notes 6, 212-217. Phan, H.T.T., Ellwood, S.R., Hane, J.K., Ford, R., Materne, M., and Oliver, R.P. (2007). Extensive macrosynteny between Medicago truncatula and Lens culinaris ssp. culinaris. Theoretical and Applied Genetics 114, 549-558. Pozarkova, D., Koblizkova, A., Roman, B., Torres, A.M., Lucretti, S., Lysak, M., Dolezel, J., and Macas, J. (2002). Development and characterization of microsatellite markers from chromosome 1-specific DNA libraries of Vicia faba. Biologia Plantarum 45, 337-345. Prioul-Gervais, S., Deniot, G., Receveur, E.M., Frankewitz, A., Fourmann, M., Rameau, C., Pilet-Nayel, M.L., and Baranger, A. (2007). Candidate genes for quantitative resistance to Mycosphaerella pinodes in pea (Pisum sativum L.). Theoretical and Applied Genetics 114, 971-984. Raji, A., Anderson, J., Kolade, O., Ugwu, C., Dixon, A., and Ingelbrecht, I. (2009). Gene-based microsatellites for cassava (Manihot esculenta Crantz): prevalence, polymorphisms, and cross-taxa utility. BMC Plant Biology 9, 118. Rao, S.C., and Northup, B.K. (2011). Growth and nutritive value of grass pea in Oklahoma. Agronomy Journal 103, 1692-1696. Schaefer, H., Hechenleitner, P., Santos-Guerra, A., Sequeira, M.M.D., Pennington, R.T., Kenicer, G., and Carine, M.A. (2012). Systematics, biogeography, and character evolution of the legume tribe Fabeae with special focus on the middle- Atlantic island lineages. BMC Evolutionary Biology 12, 250. Schuelke, M. (2000). An economic method for the fluorescent labeling of PCR fragments. Nature Biotechnology 18, 233-234. Shiferaw, E., Pè, M., Porceddu, E., and Ponnaiah, M. (2011). Exploring the genetic diversity of Ethiopian grass pea (Lathyrus sativus L.) using EST-SSR markers. Molecular Breeding 30, 789-797. Skiba, B., Ford, R., and Pang, E.C.K. (2003). Amplification and detection of polymorphic sequence-tagged sites in Lathyrus sativus. Plant Molecular Biology Reporter 21, 391-404.

65 Chapter 2 ______

Skiba, B., Ford, R., and Pang, E.C.K. (2004). Construction of a linkage map based on a Lathyrus sativus backcross population and preliminary investigation of QTLs associated with resistance to ascochyta blight. Theoretical and Applied Genetics 109, 1726-1735. Smýkal, P., Kalendar, R., Ford, R., Macas, J., and Griga, M. (2009). Evolutionary conserved lineage of Angela-family retrotransposons as a genome-wide microsatellite repeat dispersal agent. Heredity 103, 157-167. Smýkal, P., Kenicer, G., Flavell, A.J., Corander, J., Kosterin, O., Redden, R.J., Ford, R., Coyne, C.J., Maxted, N., Ambrose, M.J., and Ellis, N.T.H. (2011). Phylogeny, phylogeography and genetic diversity of the Pisum genus. Plant Genetic Resources 9, 4-18. Tavoletti, S., and Iommarini, L. (2007). Molecular marker analysis of genetic variation characterizing a grasspea (Lathyrus sativus) collection from central Italy. Plant Breeding 126, 607-611. Torres, A.M., Weeden, N.F., and Martin, A. (1993). Linkage among isozyme, RFLP and RAPD markers in Vicia faba. Theoretical and Applied Genetics 85, 937-945. Varshney, R.K., Graner, A., and Sorrells, M.E. (2005). Genic microsatellite markers in plants: features and applications. Trends in Biotechnology 23, 48-55. Varshney, R.K., Hiremath, P.J., Lekha, P., Kashiwagi, J., Balaji, J., Deokar, A.A., Vadez, V., Xiao, Y., Srinivasan, R., Gaur, P.M., Siddique, K.H., Town, C.D., and Hoisington, D.A. (2009). A comprehensive resource of drought- and salinity- responsive ESTs for gene discovery and marker development in chickpea (Cicer arietinum L.). BMC Genomics 10, 523. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2006a). Characterization of resistance to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus sativus. Plant Breeding 125, 308-310. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2007). Resistance reaction to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus cicera from Iberian origin. Genetic Resources and Crop Evolution 54, 1517-1521. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2009). Pre and posthaustorial resistance to rusts in Lathyrus cicera L. Euphytica 165, 27-34. Vaz Patto, M.C., and Rubiales, D. (2009). Identification and characterization of partial resistance to rust in a germplasm collection of Lathyrus sativus L. Plant Breeding 128, 495-500.

66 Cross-species amplification ______

Vaz Patto, M.C., Skiba, B., Pang, E.C.K., Ochatt, S.J., Lambein, F., and Rubiales, D. (2006b). Lathyrus improvement for resistance against biotic and abiotic stresses: From classical breeding to marker assisted selection. Euphytica 147, 133- 147. Wojciechowski, M.F., Lavin, M., and Sanderson, M.J. (2004). A phylogeny of legumes (Leguminosae) based on analysis of the plastid matK gene resolves many well-supported subclades within the family. American Journal of Botany 91, 1846-1862. Xu, Y., Ma, R.C., Xie, H., Liu, J.T., and Cao, M.Q. (2004). Development of SSR markers for the phylogenetic analysis of almond trees from China and the Mediterranean region. Genome 47, 1091-1104. Zhang, Y., Sledge, M.K., and Bouton, J.H. (2007). Genome mapping of white clover (Trifolium repens L.) and comparative analysis within the Trifolieae using cross-species SSR markers. Theoretical and Applied Genetics 114, 1367-1378.

67

CHAPTER 3 Allelic diversity in the transcriptomes of contrasting rust-infected genotypes of Lathyrus sativus, a lasting resource for smart breeding

The work presented in this chapter was mostly performed by Nuno Felipe Almeida (see Acknowledgements section), and corresponds to the following publication:

Almeida, N.F., Leitão, S.T., Krezdorn, N., Rotter, B., Winter, P., Rubiales, D., and Vaz Patto, M.C. (2014). Allelic diversity in the transcriptomes of contrasting rust-infected genotypes of Lathyrus sativus, a lasting resource for smart breeding. BMC Plant Biology 14, 376.

69

Grass pea response to rust ______

3.1. Abstract

Grass pea (Lathyrus sativus L.) is a valuable resource for potentially durable partial resistance to rust. To gain insight into the resistance mechanism and identify potential resistance genes, we generated the first comprehensive transcriptome assemblies from control and Uromyces pisi inoculated leafs of a susceptible and a partially rust-resistant grass pea genotype by RNA-Seq.

134,914 contigs, shared by both libraries, were used to analyse their differential expression in response to rust infection. Functional annotation grouped 60.4% of the contigs present in plant databases (37.8% of total) to 33 main functional categories, being “protein”, “RNA”, “signalling”, “transport” and “stress” the most represented. Transcription profiles revealed considerable differences in regulation of major phytohormone signalling pathways: whereas Salicylic and Abscisic Acid pathways were up-regulated in the resistant genotype, Jasmonate and Ethylene pathways were down- regulated in the susceptible one. As potential Resistance-genes we identified a mildew resistance locus O (MLO)-like gene, and MLO- related transcripts. Also, several pathogenesis-related genes were up-regulated in the resistant and exclusively down regulated in the susceptible genotype. Pathogen effectors identified in both inoculated libraries, as e.g. the rust Rtp1 transcript, may be responsible for the down-regulation of defence-related transcripts. The two genotypes contained 4,892 polymorphic contigs with SNPs unevenly distributed between different functional categories. Protein degradation (29.7%) and signalling receptor kinases (8.2%) were the most diverged, illustrating evolutionary adaptation of grass pea to the host/pathogens arms race.

71 Chapter 3 ______

The vast array of novel, resistance-related genomic information we present here provides a highly valuable resource for future smart breeding approaches in this hitherto under-researched, valuable legume crop.

3.2. Introduction

Rusts are among the most important diseases of legumes (Sillero et al., 2006) and grass pea (Lathyrus sativus L.) is not an exception (Duke, 1981; Campbell, 1997; Vaz Patto et al., 2006b). Rusts are caused by biotrophic fungi that keep infected host cells alive for their development. They form elaborate intracellular accommodation structures called haustoria, which maintain an intimate contact between fungal and plant cells over a prolonged period of time (O’Connell and Panstruga, 2006).

Rust in Lathyrus spp. is caused by Uromyces pisi (Pers.) Wint and U. viciae-fabae (Pers.) J. Schröt (Barilli et al., 2011; Barilli et al., 2012), but and in addition to Lathyrus, U. pisi infects a broad range of other legumes too (Barilli et al., 2012; Rubiales et al., 2013). Plants have developed multifaceted defence responses, many of which are induced only upon pathogen attack. These responses may include induction of pathogenesis related (PR) genes, the production of secondary metabolites (as e.g. phytoalexins), as well as the reinforcement of cell walls (Stintzi et al., 1993). Associated with these responses may be the production of reactive oxygen species (ROS) and the induction of localized cell death (the hypersensitive response, HR) (Zurbriggen et al., 2010). The induction of this basal plant defence machinery occurs upon the recognition of conserved molecules which are present in a variety of microbial species, but

72 Grass pea response to rust ______absent in the host. These pathogen associated molecular patterns (PAMPs) are molecular components highly conserved within a class of microbes, where they have essential functions for their fitness or survival (Medzhitov and Janeway Jr, 1997). These include, for example, fungal chitin, β-glucan and ergosterol. The specific virulence factors of the pathogen, known as fungal effectors, are recognized by corresponding resistance (R) genes of the host plant. Both rust- causing pathogens of Lathyrus are able to efficiently overcome R- gene based resistance (McDonald and Linde, 2002). To date, most fungal effectors identified are lineage-specific small secreted proteins (SSP) of unknown function (Stergiopoulos and de Wit, 2009; Schmidt and Panstruga, 2011). The U. viciae-fabae rust transferred protein 1 (Rtp1) was the first fungal effector visualized in the host cytoplasm and nucleus after in planta secretion by the rust fungus (Kemen et al., 2005). Rtp1 belongs to a family of cysteine protease inhibitors that are conserved in the rust species (order Pucciniales formerly known as Uredinales) (Pretsch et al., 2013).

Gene-for-gene resistance is associated with the activation of, for instance, the salicylic acid (SA)-dependent signalling pathway, leading to expression of defence-related genes like PR1, the production of ROS and finally to programmed cell death (Bari and Jones, 2009; Pieterse et al., 2012). Other phytohormones involved in plant/pathogen interaction are ethylene (ET) and jasmonates (JA). Plant defence responses appear specifically adapted to the attacking pathogen, with SA-dependent defences acting mainly against biotrophs, and JA- and ET-dependent responses acting mainly against necrotrophs (Glazebrook, 2005; O’Connell and Panstruga, 2006; Smith et al., 2009).

73 Chapter 3 ______

Grass pea is a diploid species (2n = 14) with a genome size of approx. 8.2 Gbp (Bennett and Leitch, 2012). Although grass pea is primarily self-pollinated, a 2 to 36% outcrossing rate was reported, depending on location and genotype (Rahman et al., 1995; Chowdhury and Slinkard, 1997; Gutiérrez-Marcos et al., 2006). Outcrossing is mainly driven by pollinators, and therefore can be minimized when grown in isolation (Chowdhury and Slinkard, 1997). There is a great potential for the expansion of grass pea in dry areas and zones that are becoming more drought-prone as a result of climate change (Hillocks and Maruthi, 2012). Partial resistance to U. pisi has been reported in grass pea as a clear example of prehaustorial resistance, with no associated necrosis. This resistance is due to restriction of haustoria formation accompanied by frequent early abortion of the colonies, reduction in the number of haustoria per colony and decreased intercellular growth of infecting hyphae (Vaz Patto and Rubiales, 2009). Though prehaustorial resistance is typical for non-hosts, it has also been implicated in host partial resistance (Rubiales and Niks, 1995; Sillero and Rubiales, 2002) and is common in resistance of major cool season grain legumes against rusts (Sillero et al., 2006; Rubiales et al., 2011). Additionally, resistant Lathyrus genotypes may serve as a source of new and useful genetic traits in the breeding of related major legume crops such as peas, lentils and vetches. Cross-incompatibility has been reported between pea and L. sativus, but successful fusion of Pisum sativum and L. sativus protoplasts (Durieu and Ochatt, 2000) creates new possibilities for gene transfer between these species. However, the slow progress in understanding the genetic control of important traits, such as disease resistance, in Lathyrus species hampered the development of modern cultivars or the introgression of their interesting traits into related species.

74 Grass pea response to rust ______

In economically important warm season legumes such as common bean and soybean, complete monogenically controlled resistances to rusts and associated rust resistance genes have been described together with closely linked markers for use in marker assisted backcrossing (Faleiro et al., 2004; Miklas et al., 2006; Hyten et al., 2007; Garcia et al., 2008; Rubiales et al., 2011). By contrast, most rust resistances described so far in cool season food legumes are incomplete in nature and the genetic basis of resistance is largely unknown. Although QTL mapping studies confirmed the polygenic control of resistance as e.g. in pea (Barilli et al., 2010a), faba bean (Rai et al., 2011) and chickpea (Madrid et al., 2008), no markers suitable for marker assisted selection (MAS) are available yet.

Genomic resources for grass pea are still scarce (e.g. in April 2014 the NCBI database contained only 178 EST sequences from L. sativus (Skiba et al., 2005)), and the two linkage maps existing for grass pea do not contain sufficiently informative markers to bridge between them (Vaz Patto et al., 2006b).

The advent of next-generation sequencing (NGS) technologies was an important breakthrough enabling the sensitive and quantitative high-throughput transcriptome analysis referred to as RNA-Seq (Simon et al., 2009; Metzker, 2010). RNA-Seq discriminated between microbial and host transcriptomes, during plant-microbe interactions, using original or phylogenetically related genomes as a reference for transcript annotation (Kemen et al., 2011; Fernandez et al., 2012; Tisserant et al., 2012; Westermann et al., 2012). RNA-Seq gene expression patterns provided also information on complex regulatory networks and on variations in expressed genes, such as SNPs and SSRs, in an increasing number of non-

75 Chapter 3 ______model plants (Strickler et al., 2012) and thus may be well suited to overcome the bottleneck of lacking genomic resources in Lathyrus.

Here we employed RNA-Seq to study the response of L. sativus to U. pisi infection. We used MapMan and metabolic pathway analyses to interpret the results and assessed allelic diversity in transcripts as a source for genic markers for future (comparative) mapping studies. In addition, the expression of a set of selected genes was measured by RT-qPCR to validate the RNA-Seq results.

To our knowledge, this is the first study on the global expression profiling of genes in grass pea/pathogen interaction using NGS. Our results will assist the elucidation of pathways and genes associated with resistance to rust in grass pea and related species. This approach may represent one of the initial steps towards the development of effective strategies for resistance breeding against such a quickly evolving pathogen.

3.3. Material and methods

3.3.1. Plant and fungal material, inoculation and RNA isolation

The two L. sativus genotypes, BGE015746 and BGE024709 analysed in the present work were kindly provided by the Plant Genetic Resources Centre (CRF-INIA), Madrid, Spain. Seeds were multiplied in insect proof cages in order to minimize outcrossing. Evaluation for their resistance against U. pisi demonstrated that BGE024709 is susceptible to rust, whereas BGE015746 displays partial resistance (Vaz Patto and Rubiales, 2009). Upon infection, both genotypes present well-formed pustules, with no associated chlorosis or necrosis. They contrast, however, in disease severity (DS), i.e. the percentage of leaf area covered by the fungus. Whereas

76 Grass pea response to rust ______the partial resistant genotype has a DS=9%, the susceptible one has a DS=30%.

The U. pisi monosporic isolate UpCo-01 from the fungal collection of the Institute for Sustainable Agriculture-CSIC (Córdoba, Spain) was used for the experiment. Inoculum was multiplied on plants of the susceptible P. sativum cv. Messire before use.

Twenty-four plants per each genotype and treatment (inoculated/control) were used. Two-week-old L. sativus seedlings were inoculated by dusting all the plants at the same time with 2 mg of spores per plant, diluted in pure talk (1:10), with the help of a small manual dusting device in a complete random experiment. Inoculated and control plants were incubated for 24 h at 20 ºC, in complete darkness, and 100% relative humidity, then transferred to a growth chamber and kept at 20 ± 2 ºC under 14h light (150 μmol m−2 s−1) and 10h dark.

RNA was extracted from inoculated and non-inoculated fresh leaves collected 37 hours after inoculation. The material was immediately frozen in liquid nitrogen and stored at -80 °C. RNA was isolated using the GeneJET Plant RNA Purification Mini Kit (Thermo Scientific, Vilnius, Lithuania), according to the manufacturer’s instructions. The extracted RNA was treated with Turbo DNase I (Ambion, Austin, TX, USA), and RNA quantification was carried out using the NanoDrop device (Thermo Scientific, Passau, Germany).

3.3.2. Sequencing and quantification

For each of the 4 combinations, genotype and treatment (BGE015746 control, BGE015746 inoculated, BGE024709 control and BGE024709 inoculated) total RNA from 24 plants was extracted

77 Chapter 3 ______and pooled in equal amounts for sequencing. Three RNA-Seq libraries (one for each genotype and one reference assembly, including all genotypes and treatments) were generated by GenXPro GmbH, Germany, using a proprietary protocol. In short, mRNA was captured from 20 µg of total RNA using Oligo dT(25) beads (Dynabeads; life Technologies). The purified mRNA was randomly fragmented in a Zn2+ solution to obtain approximately 250 bp long RNA fragments. cDNA was synthesized by reverse transcription starting from 6(N) random hexamer oligonucleotides followed by second strand synthesis. Barcoded Y-adapters were ligated to the cDNA and the library was amplified with 10 cycles of PCR. The libraries were sequenced on an Illumina Hiseq2000 machine. After Illumina paired-end sequencing, raw sequence reads were passed through quality filtering, thereby also removing sequencing adapter primers and cDNA synthesis primers. All high-quality reads were assembled using the Trinity RNA-Seq de novo assembly (Version: trinityrnaseq_r2011-11-26). In order to minimize the redundancy, CAP3 software (Huang and Madan, 1999) was also used with overlap length cutoff of 30 bp and overlap percent identity cutoff of 75%. Redundancy was tested using the clustering algorithm UCLUST ((Edgar, 2010), available at http://drive5.com/usearch/ manual/uclust_algo.html). The resulting contigs were annotated via BLASTX to publically available databases (ftp://ftp.ncbi.nlm.nih.gov/blast/db/FASTA/nr.gz, nr, plants only). To identify fungal transcripts, an additional BLASTX to public fungal databases (http://www.ebi.ac.uk/uniprot, UniProtKB/Swiss-Prot and UniProtKB/TrEMBL) was performed. The sequenced reads were mapped with novoalign software (V2.07.14; http://www.novocraft.com/) to the own assembled contigs. RPKM was calculated as the normalized transcript expression value (Marioni et

78 Grass pea response to rust ______al., 2008). Our obtained counts were subsequently passed through DEGSeq to calculate the differential gene expression (R package version 1.16.0) (Wang et al., 2010).

3.3.3. SNP detection

SNPs were discovered between the two L. sativus genotypes using JointSNVMix (Roth et al., 2012). The mappings from the transcriptome analysis were also analysed by JointSNVMix and the output was furthermore processed by GenXPro’s in-house software to detect SNPs discriminating the bulks. SNP calling was performed taking in account only the inoculated samples. A minimum coverage of 15 reads in each genotype in the inoculated condition was needed to call a SNP. Polymorphic contigs and their respective SNPs are listed in Additional File 3.1.

3.3.4. EST-SSR development and genotyping

EST-SSRs were selected in silico by identifying the polymorphic SSRs between the two L. sativus genotypes. Identification of the SSRs was done using Phobos plug-in (Mayer, 2010) for the Geneious software (Drummond et al., 2011), using as search parameters, perfect SSRs with a repeat unit length of two to six nucleotides. Length polymorphisms were manually identified by aligning SSR-containing contigs of one genotype against the whole library of the other genotype. Primers were designed using Primer3 plug-in (Untergasser et al., 2012) for the Geneious software, using as parameters a melting temperature from 59 to 63ºC, a GC content of 50 to 60% and a primer size ranging from 18 to 24 nucleotides. The developed EST-SSR markers are listed in Additional File 3.2.

79 Chapter 3 ______

PCR reactions for the EST-SSRs genotyping were conducted using the M13 tail labelling strategy described by Schuelke (2000) in a total volume of 10 µl containing 10 ng of template DNA, 0.04 µM of M13(-21) tagged forward primer, 0.16 µM of IRD700 or IRD 800 M13(-21) and 0.16 µM of reverse primer, 0.2 mM of each dNTP, 1.5 mM of MgCl2, and 0.2 unit of Taq DNA polymerase (Promega, Madison, USA). The amplification reaction consisted of a denaturing step of 5 min at 94ºC, followed by 30 cycles of 30 s at 94ºC, 45 s at 56ºC, 45 s at 72ºC, and 8 cycles of 30 s at 94ºC, 45 s at 53ºC, 45 s at 72ºC. The reaction was terminated at 72ºC for 10 min.

SSR fragments were resolved with 6.5% polyacrylamide gel using a LI-COR 4300 DNA Analyzer (Lincoln, NE, USA).

3.3.5. Quantitative RT-PCR assay

1 μg of total RNA from each of three randomly chosen plants per genotype, per treatment (inoculated/control), was reverse transcribed in duplicates, using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) following manufacturer’s instructions. Three independent reverse- transcription reactions (RT) were performed for each cDNA sample in a total of nine samples per genotype, per treatment. For all genes studied, the product of each of these reactions was analysed in technical duplicates, in a total of six technical replicates per treatment. RT-qPCR reactions were performed with a iQ™5 Real- Time PCR Detection System (Bio-Rad, Munich, Germany). Primers were designed using the Primer3 software (Untergasser et al., 2012). Primer sequences can be found in Additional File 3.3. For data analysis, the Genex software package (MultiD, Goteborg, Sweden),

80 Grass pea response to rust ______including the geNorm software (Vandesompele et al., 2002) was used.

3.3.6. Contig annotation and data analysis

In order to classify the contigs into functional categories, the Mercator pipeline for automated sequence annotation ((Lohse et al., 2014), available at http://mapman.gabipd.org/web/guest/app/ mercator) was employed. The mapping file was created excluding contigs without BLAST hit in previous analyses and accessing the following, manually curated databases: Arabidopsis TAIR proteins (release 10), SwissProt/UniProt Plant Proteins (PPAP), TIGR5 rice proteins (ORYZA), Clusters of orthologous eukaryotic genes database (KOG), Conserved domain database (CDD) and InterPro scan (IPR). The Mercator mapping file was then employed for pathway analysis by the MapMan software ((Thimm et al., 2004), available at http://mapman.gabipd.org/web/guest/mapman).

Differentially expressed contigs were identified by comparing their expression in leaves of the resistant genotype BGE015746, control vs. inoculated, and of the susceptible BGE024709, control vs. inoculated, using DEGSeq (Wang et al., 2010). In cases where a particular transcript reacted in the same way in both genotypes, the total transcript count before and after inoculation was compared, allowing the identification of basal genotypic differences between the two genotypes.

3.3.7. Availability of supporting data

The raw RNA-Seq data supporting the result of this article is available in the Sequence Read Archive (SRA), with accession

81 Chapter 3 ______numbers SRS686331, SRS687370, SRS687371 and SRS687373. This Transcriptome Shotgun Assembly (TSA) project has been deposited at DDBJ/EMBL/GenBank under the accession GBSS00000000. The version described in this paper is the first version, GBSS01000000.

3.4. Results

3.4.1. Contigs from the RNA-Seq transcriptomes of resistant and susceptible L. sativus genotypes

The RNA-Seq libraries from control and inoculated leafs from the resistant genotype BGE015746 were united prior to assembly to generate a comprehensive data set enabling the generation of contigs of maximum length. They included 46,994,629 reads which were assembled into 105,288 contigs, ranging in size from 150 to 13,929 bp, with a mean contig length of 544 bp. The respective united library from the susceptible genotype BGE024709 comprised 72,566,465 reads which assembled in 119,870 contigs, with a size range of 150 to 15,658 bp and a mean contig length of 524 bp.

A reference assembly using both genotypes and treatments assembled in 134,914 contigs, ranging in size from 150 to 13,916 bp, with a mean contig length of 501 bp. The mapping and quantification of both genotypes’ libraries to the reference assembly allowed the analysis of their differential expression in response to U. pisi infection. 9,501 contigs were unique to the resistant and 15,645 contigs were unique to the susceptible genotype. Redundancy of the reference assembly was checked using the clustering algorithm UCLUST, identifying only 49 (0.036%) transcripts with identity higher than 95%.

82 Grass pea response to rust ______

This Transcriptome Shotgun Assembly project has been deposited at DDBJ/EMBL/GenBank under the accession GBSS00000000. The version described in this paper is the first version, GBSS01000000.

3.4.2. RNA-Seq validation by quantitative RT-PCR assay

To validate the RNA-Seq results, expression levels of a set of 9 selected genes were analysed by RT-qPCR. Genes were selected by their level of expression and transcript count, in order to represent a broad range of expression profiles. Further, the number of their transcripts differed between inoculated and control samples by log2 ratios ranging from -6.39 to 4.70 at q-values < 0.05. Their read count numbers were generally higher than 100, with exception of contig a45744;151, “mitochondrial chaperone BCS1”, with 2 counts in the resistant control and 36 counts in the resistant inoculated line, and contig a32859;123 “seed maturation protein”, with 3 counts in the susceptible inoculated line and 104 counts in the resistant inoculated line (Table 3.1). The best housekeeping genes for normalization suggested by the geNorm software were, for the resistant genotype samples, “β-tubulin” (a6507;507) and “photosystem I P700 apoprotein A2” (a160;902), and “O-methyltransferase” (a5102;390), for the susceptible genotype. A good correlation (R=0.82 for the resistant and R=0.80 for the susceptible genotypes) was observed between the log2 fold changes measured by RNA-Seq and RT-qPCR (Figure 3.1).

83 Chapter 3 ______

3.4.3. Differential gene expression in resistant and susceptible L. sativus genotypes during infection

Differentially expressed contigs were grouped by expression patterns based on up- or down-regulation (log2 ≥ 2 or log2 ≤ -2; respectively, q-value ≤ 0.05) after inoculation. Within each expression pattern group, comparisons were performed between genotypes. Expression patterns were grouped in eight response types, according to their up- or down-regulation, in susceptible and resistant genotypes, respectively. The number of contigs and description of each group is summarized in Table 3.2. Most representative groups are group F (contigs down regulated in both genotypes) and H (contigs down-regulated only in the resistant genotype) with 2,516 and 1,606 contigs respectively, followed by group A that includes 814 contigs up-regulated in both resistant and susceptible genotypes upon infection. A detailed list with all the identified contigs, their description and expression pattern groups can be found in Additional File 3.4.

84 ______Table 3.1 - Log2 fold expression results for RNA-Seq and RT-qPCR experiments. RPKM: reads per kilobase per million

BGE015746 BGE024709 Reference inoculated/ inoculated/ inoculated/ inoculated/ assembly BLAST hit control control inoculated inoculated DEGSeq control control inoculated inoculated DEGSeq control control qPCR control control qPCR contig counts RPKM counts RPKM q-value counts RPKM counts RPKM q-value DEGSeq (log2) (log2) DEGSeq (log2) (log2)

Chromodomain a1310;251 helicase DNA- 4063 0.0362 6436 0.0396 -0.53 1.0E-77 0.83 4149 0.0287 8009 0.0379 -0.33 1.7E-33 5.51 binding protein

Type IIB calcium a15017;192 850 0.0135 1855 0.0204 -0.07 3.2E-01 1.90 1165 0.0144 4635 0.0391 0.72 6.5E-64 3.38 ATPase

Amino acid a19532;154 1965 0.0312 1328 0.0145 -1.76 8.4E-266 1.73 2045 0.0251 2224 0.0187 -1.15 3.2E-151 0.42 transporter

Hypothetical protein a22579;158 1329 0.0592 901 0.0277 -1.76 2.2E-179 -0.39 602 0.0208 245 0.0058 -2.57 3.0E-135 2.61 MTR 2g062700

a2401;404 Lectin 776 0.0415 1061 0.0392 -0.75 1.5E-27 -1.35 813 0.0337 4283 0.1214 -1.12 6.8E-125 -1.92

Seed maturation a32859;123 276 0.0438 185 0.0202 -1.77 1.9E-38 -3.74 104 0.0128 3 0.0003 -6.39 4.9E-39 -1.81 protein

Mitochondrial a45744;151 2 0.0004 36 0.0051 2.97 7.8E-05 6.29 7 0.0011 440 0.0480 4.70 1.2E-65 8.35 chaperone BCS1

Alpha-galactosidase a5330;269 1092 0.0373 1443 0.0341 -0.79 8.9E-43 -0.60 1203 0.0319 2387 0.0433 -0.29 2.8E-08 1.91

1 Grass pea response to rust

GDSL a6560;334 esterase/lipase 1981 0.0904 1917 0.0604 -1.24 3.3E-160 0.00 2164 0.0765 1101 0.0266 -2.25 0.0E+00 0.39 EXL3 8 5

Chapter 3 ______

Figure 3.1 - Correlation between RNA-Seq and RT-qPCR. The relative expression levels obtained by RNA-Seq using DEGSeq and by RT-qPCR using the ΔΔCt method. Pearson’s correlation coefficient (R) between relative expression levels is shown above the trend line.

As depicted in Figure 3.2, from the 134,914 contigs that could be identified and quantified, 68,889 were shared among all libraries. Of these, 974 contigs were up-regulated and 5,203 contigs down- regulated in the resistant genotype BGE015746 and 772 contigs up- and 4,617 down-regulated in the susceptible genotype BGE024709. Furthermore, from the 5,807 contigs only present in the resistant genotype’s libraries, 132 were up- and 485 down-regulated (inoculated vs. control). From the 7,938 contigs only found in the susceptible genotype’s libraries, 134 were up- and 689 down- regulated.

86 Grass pea response to rust ______

Table 3.2 - Classification of contigs according to their differential expression in the susceptible and resistant genotype upon infection with U. pisi. Up regulated: (log2 >= 2; q- value ≤ 0.05); Down-regulated: (log2 ≤ -2; q-value ≤ 0.05); higher in Susceptible: (log2 fold change between all resistant and susceptible genotype contigs <= -2; q-value ≤ 0.05); higher in Resistant: (log2 fold change between all resistant and susceptible genotype contigs >= 2; q- value ≤ 0.05)

Expression # of Feature pattern group contigs Up-regulated in Resistant A 814 Up-regulated in Susceptible Up-regulated in Resistant B 32 Up-regulated in Susceptible, higher in Susceptible Up-regulated in Resistant, higher in Resistant C 56 Up-regulated in Susceptible

D Up-regulated in Susceptible 319

E Up-regulated in Resistant 576

Down-regulated in Resistant F 2,516 Down-regulated in Susceptible

G Down-regulated in Susceptible 548

H Down-regulated in Resistant 1,606

Total 134,914

3.4.4. Annotation

From the 134,914 contigs detected in all libraries, 50,937 (37.75%) contigs could be matched via BLAST to entries in plant databases and 961 (0.71%) matched only to fungal databases. The latter contigs were present only in the inoculated libraries. Also, 4,558 contigs were absent in control samples and found exclusively in fungal databases, or with a higher bit-score in fungal databases than in plant databases and thus, most probably correspond to U. pisi sequences.

87 Chapter 3 ______

Figure 3.2 - Venn diagram of the number of unique and shared contigs between the two genotypes and its expression. In black boxes the number of up (log2 fold ≥ 2) and down (log2 fold ≤ -2) regulated contigs in the inoculated condition versus control. Resistant genotype: BGE015746, susceptible genotype: BGE024709.

As indicated in Figure 3.3, BLAST produced hits mainly to other legume species with frequencies in the order Medicago truncatula (26,728; 19.81%), Glycine max (11,436; 8.48%), P. sativum (1,538; 1.14%) and Lotus japonicus (921; 0.68%). Vitis vinifera (2,409; 1.79%), Populus trichocarpa (656; 0.49%) and the model Arabidopsis thaliana (607; 0.45%) were the best matching non- legume species. BLAST hits from L. sativus comprised only 0.02% (33 contigs) of the total illustrating the scarcity of Lathyrus entries in the data bases.

88 Grass pea response to rust ______

Figure 3.3 - Number of contigs that could be BLASTed to different plant species.

From the 4,558 contigs that were absent in control samples and found exclusively in fungal databases, or with a higher bit-score in fungal databases, 20 contigs from the 49 accessions described in UniProtKB/Swiss-Prot and UniProtKB/TrEMBL as U. viciae-fabae, were identified (see list in Additional File 3.5). None of these 20 contigs were significantly differentially expressed between the two inoculated genotypes. For example, among the eight contigs out of the 20 without a plant database hit, five were homologous to “invertase 1”, and the three others to “rust transferred protein – Rtp1”, “amino acid transporter” and “putative permease”. Six other contigs absent in control samples and found exclusively in fungal databases or with a higher bit-score in fungal databases, were homologous to housekeeping genes that can be found throughout different kingdoms (three “tubulin beta chain”, two “succinate dehydrogenase” and one “plasma membrane (H+) ATPase”.

89 Chapter 3 ______

Functional annotation of the contigs via Mercator and MapMan, depicted in Figure 3.4, grouped 60.4 % of them into 33 main functional categories, of which the categories “protein” (11.0%), “RNA” (8.0%), “signalling” (6.7%), “transport” (5.4%) and “stress” (4.2%) were most crowded. A total of 39.4% could not be assigned to any functional category.

Analysis of functional categories, within each expression pattern group, identified differences among the functions present within each group. Comparisons were also performed among the different expression profiles in each category (Figure 3.5). Transcripts included in the functional categories “stress” and “protein” were present at a higher percentage in up-regulated expression pattern groups (“stress” in A, B and C; “protein” in A, B and E), while the functional category “cell” was present at higher percentage in down- regulated expression pattern groups (F, G and H). The most prominent functional category in group C contigs up-regulated in both genotypes, with a higher expression in the resistant genotype was “cell wall”. “Lipid metabolism” and “DNA” were also over-represented. However, also the down-regulated groups F, G, and H contained a considerable number of contigs from the “cell wall” category. In group B, joining contigs up-regulated in both genotypes with a higher expression in the susceptible genotype, the categories “secondary metabolism” and “hormone metabolism” were over-represented. Interestingly, the functional category “signalling” was over- represented in contigs up- regulated only in the susceptible genotype, as in group D.

90 ______Grass pea response to rust

Figure 3.4 - Percentage of contigs assigned in each main functional category. 91

Chapter 3 ______

3.4.5. Biotic stress related proteins

In order to restrict the number of analysed contigs to the ones probably more directly related to resistance, we focused mostly on contigs up-regulated at a higher ratio, or exclusively, in the resistant genotype (groups C and E), contigs exclusively down-regulated in the susceptible genotype (group G) and contigs exclusively down- regulated in the resistant genotype (group H).

From the subcategory “stress.biotic”, two contigs in group E corresponded to the well- studied mildew resistance locus O (MLO) gene which was first identified in barley, conferring resistance to powdery mildew (Jørgensen, 1992). Also, from a total of 25 “MLO- like” contigs, 12 were differentially expressed. Two of these (a116583;40 and a25504;132) were down-regulated in the resistant genotype (group H). These might be related to MLO susceptibility genes, as reported by several previous studies (Kim and Hwang, 2012; Zheng et al., 2013; McGrann et al., 2014). Interestingly, in the susceptible genotype, one “PREDICTED: beta glucosidase 12-like”, identified by Mercator as “PENETRATION 2”, required for MLO- mediated resistance and belonging to the functional category “secondary metabolism”, was down-regulated (group G). Group G also contained one “acidic endochitinase” and two leucine-rich repeat (LRR) proteins, one TIR-NBS-LRR and one containing LRR and NB- ARC domains. In group C, a pathogenesis related protein 1 (PR-1) contig was identified.

92 ______Figure 3.5 - Percentage of contigs assigned in each functional category for each expression pattern group. A - contigs up-regulated in both resistant and susceptible genotypes similarly; B - contigs up-regulated in both genotypes, with a higher expression in the susceptible genotype; C - contigs up-regulated in both genotypes, with a higher expres- sion in the resistant genotype; D - contigs up-regulated only in the susceptible genotype; E - contigs up-regulated only in the resistant genotype; F - contigs down-regulated in both resistant and susceptible genotypes similarly; Grass pea response to rust G - contigs down-regulated only in the susceptible genotype; H - contigs down-regulated only in the resistant genotype. 93

Chapter 3 ______

The subcategory “stress.abiotic” contained, i.a., genes involved in response to heat that also respond to biotic stresses. For example, in group C and E, we identified one “DNAJ heat shock protein” in both groups, three “heat shock protein 70 family” (group E) and one “18.1 kDa class 1 heat shock protein” (group E). Group G, however, contained one “DNAJ homolog subfamily B member” and one “double Clp-N motif-containing P-loop nucleoside triphosphate hydrolases superfamily protein”.

Several contigs related to secondary metabolism were exclusively up-regulated in the resistant genotype (group E). These comprised a “reticuline oxidase-like protein” involved in alkaloid biosynthesis, an “isoflavone 2’hydroxylase”, functioning in the isoflavonoid biosynthesis pathway, a “dihydroflavonol-4-reductase”, with roles in the flavonoid and brassinosteroid metabolic pathway and an “AMP-dependent CoA ligase”, acting in the JA and lignin biosynthesis pathways. In group G, 17 contigs were related to secondary metabolism including four involved in the flavonoid pathway, two in the isoprenoid/terpenoid pathway and one “WAX 2- like” involved in wax biosynthesis.

PTI (pathogen-associated molecular pattern triggered immunity) relies on an efficient signalling network in order to control the infection (Nicaise et al., 2009). Receptor kinases are important for the plant’s pathogen recognition and their expression may be constitutively expressed or up-regulated in resistant genotypes or down-regulated in susceptible genotypes in response to effectors from the pathogen. Receptor kinases and kinases exclusively up- regulated in the resistant genotype and contained in group E may be part of such signalling cascades. These included one protein kinase with thaumatin (PR-5) domain, six “DUF 26”, one “CRINKLY4”, one

94 Grass pea response to rust ______

“FERONIA receptor like kinase”, and also a MAP kinase “MAPKKK5” and a G-protein “zinc finger (Ran-binding) family protein”. In contrast, the down-regulation of such transcripts in the susceptible genotype (group G) may contribute to susceptibility. Here we identified three “DUF 26”, three LRR (“NIK1”, “RKF1” and “PXY”), two G-proteins (“guanine nucleotide-binding protein” and “dynamin-related protein 1E-like”), two MAP kinases (“PAS domain-containing protein tyrosine kinase family protein”) and three genes involved in calcium signalling (“calcium-transporting ATPase”, “calmodulin-binding heat-shock protein” and “calmodulin-domain protein kinase 9”). Interestingly, calmodulin also plays a role in the MLO response, where the lack of a calmodulin binding site decreases its defence response (Kim et al., 2002).

The “cell wall” category contained seven cellulose synthase contigs: one in group E “IRREGULAR XYLEM 3 (IRX3)” and four in group C (three “IRX1” and one “CESA1”). In group G, we identified two cellulose synthase “IRX14” and two “pectinesterase inhibitor” contigs.

From the genes normally associated with defence response, only one “endo-beta-1 3-glucanase” was identified in group C, while two others “endo-beta-1 3-glucanase” were detected in group G. Also in group G, we identified two “peroxidase” and two “glutathione S- transferase” genes.

95 Chapter 3 ______

3.4.6. SNPs in resistance pathways

In the 68,889 contigs present in both the susceptible and the resistant genotypes, we identified 2,634 contigs containing Single Nucleotide Polymorphisms (SNPs) discriminating between their respective alleles. The number of SNPs in functional (MapMan) categories varied considerably. The categories “RNA regulation of transcription” (9.5%) and “protein.degradation” (8.9%) contained by far the most SNPs, followed by the protein-related categories “protein.postranslational modification” (4.3%) and “protein.synthesis” (3.2%). Other categories including the most SNP-containing contigs were “signalling.receptor kinases” (2.5%), “protein.targeting” (2.4%) and the stress related categories, “stress.biotic” (1.8%) and “stress.abiotic” (1.6%) (Figure 3.6).

3.4.7. EST-SRR development

200 EST-SSR potential polymorphic markers between the two genotypes were designed. EST-SSRs were identified by the Phobos software (Mayer, 2010), using as search parameters, perfect SSRs with a repeat unit lenght of two to six nucleotides. Polymorphisms between the resistant and susceptible genotypes were manually identified and flanked by primer pair using the Primer3 software (Untergasser et al., 2012). To validate the EST-SSR sequences, 40 primer pairs were randomly selected for PCR amplification to confirm the presence of size polymorphism between the two accessions. PCR reactions were conducted twice in order to confirm the results. From the total 40 EST-SSR tested, 25 (62.5%) primer pairs successfully amplified polymorphic fragments between the two accessions. 6 (15.0%) primer pairs

96 ______Grass pea response to rust

Figure 3.6 - Percentage of contigs containing SNPs between the resistant and susceptible genotypes in each Mercator mapping functional subcate- gory. FA: fatty acid; metabol.: metabolism; misc.: miscellaneous; PS: photosynthesis; red.: redox 9 7

Chapter 3 ______amplified monomorphic fragments and 5 (12.5%) produce a very complex pattern. The remaining 4 (10.0%) primer pairs were not able to produce any fragments.

3.5. Discussion

Lathyrus spp. is a potential source of resistance to several pathogens (Vaz Patto et al., 2006a; Hillocks and Maruthi, 2012) and especially L. sativus provides resistance to several fungal and bacterial diseases (Skiba et al., 2005; Vaz Patto et al., 2006a; Vaz Patto and Rubiales, 2009; Martín-Sanz et al., 2012). However, the lack of genetic and/or genomic information was a barrier to further identify resistance-related genes and to use them in breeding.

In the present study we therefore attempted to improve this unfavourable situation by identifying ESTs and SNPs, potentially involved in resistance, that may be used in future smart breeding approaches. We describe for the first time a high-throughput transcriptome assembly of grass pea/pathogen interaction, using genotypes contrasting in response to rust infection, to unravel the involved partial resistance mechanism and associated resistant genes.

Our study has identified a large number of differentially expressed genes corresponding to biological categories that are thought to be most relevant in grass pea response to rust. A limitation of our study is the fact that only a single pooled sample was investigated for each genotype and condition. Although the biological variance could not be assessed in the bulked approach, the large number of individual samples in the pool is likely to level out many of possible outliers. Nevertheless, the validation of twelve genes by RT-

98 Grass pea response to rust ______qPCR, using three biological replicates, provided a good correlation with RNA-Seq results.

Another motive that could also be influencing our results is that we used different cDNA synthesis primers, oligo(dT) for the RNA- Seq and poly(A) for RT-qPCR, what might yield different quantities of poly-adenylated and non-adenylated transcripts.

Our study was severely hampered by the low number of annotated sequences, which is due to the lack of a reference genomic sequence for Lathyrus. Nevertheless, we could annotate between 34% and 46% of differentially expressed contigs to hits in plant databases, depending on the genotype and the infection status of the plants. We further developed new gene-based molecular tools as e.g. expressed sequence tags, gene-based simple sequence repeats (EST-SSR) and SNP-based markers. Moreover, we identified a number of U. pisi effectors in the infected tissues though the overall low number of observed fungal transcripts probably reflects the low quantity of fungal structures in early-infected leaves (Hacquard et al., 2011). Thus, our present study will help to overcome the problems we encountered in previous work, where the transfer of molecular markers from close related species had a very low rate of success (18% for pea EST-SSRs and 6% for pea genomic SSRs, (Almeida et al., 2014)) Therefore, the present RNA-Seq libraries will boost the availability of specific EST-SSRs and SNP-based markers that will be equally important for future development of more effective grass pea resistance breeding approaches.

The high amplification rate of the developed EST-SSRs validates the quality of the RNA-Seq data. The few primers that failed to produce amplification products or produced amplicons with an unexpected pattern may be caused by the location of the respective

99 Chapter 3 ______primers across splice regions or the presence of a large intron, since genomic regions are absent from cDNA. In addition also primers could be derived from chimeric cDNA clones (Varshney et al., 2005)].

Besides the novel markers, the deep insights into pathogenesis-related mechanisms provided by this study are of particular interest. The most interesting pathogenesis-related protein that we identified, the “MLO-like protein” is involved in signalling in response to biotic stress. MLO was described for the first time in barley, where its loss of function conferred partial resistance to powdery mildew by inducing the thickening of the cell wall at fungal penetration sites (Jørgensen, 1992). Two “MLO-like” contigs were down-regulated exclusively in the resistant genotype (group E), and perhaps related to this, we identified cellulose biosynthesis genes. The exclusively resistance-up-regulated group E contained one “IRREGULAR XYLEM 3” (IRX3) gene and three “IRX1”. Additionally, one “cellulose synthase 1” (CESA1) was stronger up-regulated in the resistant genotype than in the susceptible one (group C). Consistent with the assumed importance of MLO signalling for rust resistance, some genes important for MLO function as e.g. “calmodulin”, involved in calcium signalling as a prerequisite for MLO function (Kim et al., 2002), were down-regulated in the susceptible genotype (group G). Previously, several MLO orthologs were already demonstrated to function as susceptibility genes (Kim and Hwang, 2012; Zheng et al., 2013; McGrann et al., 2014). Therefore, we consider these two MLO- like transcripts (a116583;40 and a25504;132) as good potential candidate susceptibility genes. In order to confirm this assumption, callose deposition, as a potentially durable resistance mechanism against rusts, should be further investigated in rust-resistant and susceptible grass pea genotypes.

100 Grass pea response to rust ______

Plant responses to biotic stressors are, i.a., controlled by phytohormones as e.g. salicylic acid (SA), abscisic acid (ABA), jasmonates (JA) and ethylene (ET). Differences in expression of hormone-related genes of the susceptible and resistant genotype, in response to the pathogen, also occurred in our gene expression patterns. For example, plant resistance to biotrophic pathogens is mainly controlled by the SA pathway (Bari and Jones, 2009) and the importance of SA in the induction of systemic acquired resistance in legumes against rust fungi has been reported (Barilli et al., 2010b; Sillero et al., 2012). In our study an inducer of the SA pathway, the “ethylene response factor 5” (ERF5) gene, which at the same time inhibits the JA and ET biosynthesis pathways (Son et al., 2011), was exclusively up-regulated in the resistance genotype (group E), whereas two Apetala2/Ethylene Responsive Factor (AP2/ERF) transcription factor genes, important for the regulation of defence responses (Gutterson and Reuber, 2004), were down- regulated in the susceptible genotype (group G).

ABA regulates defence responses through its effects on callose deposition and production of ROS intermediates (Bari and Jones, 2009), activating also stomata closure as a barrier against pathogen infection (Melotto et al., 2006). In the resistant genotype, the transcript for “9-cis-epoxycarotenoid dioxygenase 2”, a key regulator of ABA biosynthesis in response to drought (Qin and Zeevaart, 1999), and involved in the crosstalk between ABA and SA signalling in plant-pathogen interactions (Cao et al., 2011), was up- regulated (group E), whereas several transcripts engaged in ABA, auxin and JA signalling, were down-regulated in the susceptible genotype (group G). This is consistent with a susceptible response to a biotroph attack (Bari and Jones, 2009).

101 Chapter 3 ______

The image emerging from the transcription profiles, of the resistant and susceptible genotype, further highlights that pathogenesis related (PR) proteins are key players in Lathyrus-rust interactions since several PR genes were mainly up-regulated in the resistant genotype, after inoculation. Among these were two chitinases (PR-3 and PR-9) involved in the degradation of the fungal cell wall [9] and a thaumatin (PR-5) gene, which causes an increase of the permeability of fungal membranes by pore-forming mechanisms (Selitrennikoff, 2001). In group E, we found a “pathogenesis related protein 1” (PR-1) and a “protein kinase-coding resistance protein”, a receptor kinase with a thaumatin domain (PR5K), presumably involved in thaumatin signaling and described previously as delaying infection (Guo et al., 2003). Another important PR-gene, an “acidic endochitinase” (PR3), was down regulated exclusively in the susceptible genotype. Genes involved in secondary metabolism were also detected. Legumes utilize flavonoids, notably isoflavones and isoflavanones, for defence against pathogens and as signalling molecules, with a number of phenylpropanoids having antimicrobial activity and restricting pathogen growth and disease symptoms (Rojas-Molina et al., 2007). In group G, we identified a “reticuline oxidase-like protein”, up regulated in non- race-specific resistance to stripe rust in wheat (Chen et al., 2013), an “isoflavone 2'-hydroxylase” from the isoflavonoid pathway (Liu et al., 2003) and a “dihydroflavonol-4-reductase” catalysing the first enzymatic step in anthocyanin biosynthesis, in the flavonoid pathway (Shimada et al., 2004).

Also, exclusively down regulated in the susceptible genotype, we found some genes important for defence response within the miscellaneous category, like “endo-beta-1 3-glucanase”, “glutathione

102 Grass pea response to rust ______

S-transferase” (GST) and “peroxidase”. In plants, beta-glycosidases serve a number of diverse and important functions, including bioactivation of defence compounds, cell wall degradation in endosperm during seed germination, activation of phytohormones, and lignifications (Morant et al., 2008). GSTs are detoxification- related proteins, protecting cells from secondary metabolites produced in response to pathogen attack, including peroxidases (Marrs, 1996). Finally, peroxidases function as resistance factors against invading fungi, inhibiting hyphal elongation, and when H2O2 is present, causing oxidative burst (Ghosh, 2006).

Effectors are expected to be excellent targets for the control of pathogens, but, unlike effectors from some other plant pathogens, relatively little is known about rust effectors (Link et al., 2014). In this study, several unigenes were identified in fungal databases. The most known rust effector identified was “rust transferred protein 1” (Rtp1). This effector aggregates into amyloid-like filaments in vitro (Kemen et al., 2013). Immunoelectron microscopy localized this effector to the extrahaustorial matrix protuberances extending into the host cytoplasm, although the exact role for this protein remains to be discovered (Giraldo and Valent, 2013). Other Uromyces effectors identified in this study were “succinate dehydrogenase”, “invertase” and “permease”. From the total potential rust transcripts identified, a selection of effector proteins could be used as probes to identify the target host proteins as a first step in the development of effector- driven legume breeding, maximizing the durability of resistance against the quickly evolving rust pathogens (Vleeshouwers et al., 2011).

From the DE contigs obtained in the present study, 2,634 presented SNPs between the resistant and the susceptible lines. The

103 Chapter 3 ______

MapMan software aided in the functional categorisation of SNPs, revealing that the categories “RNA regulation of transcription” (9.5%) and “protein.degradation” (8.9%) contained by far the most SNPs. Within these categories, ubiquitins were most polymorphic (5.6%). Ubiquitins tag proteins for proteasome degradation and play a central role in signalling pathways (Marino et al., 2012). Especially ubiquitin “E3 RING” and “SCF F-BOX” contigs contained a large number of SNPs (1.7% and 1.8% respectively). E3 RING and SCF F-BOX proteins are involved in several aspects of plant immunity ranging from pathogen recognition to both PTI to effector-triggered immunity (ETI). From the differentially expressed contigs identified as containing SNPs, we found one E3 RING, “PREDICTED: RING-H2 finger protein ATL2-like”, down regulated in the susceptible genotype. Four other functional categories, “RNA.regulation of transcription” (9.5%), “protein.postranslational modification” (4.3%), "protein.synthesis" (3.2%) and "signalling.receptor kinases" (2.5%), also contained significant numbers of polymorphisms. Especially the “signalling.receptor kinases” category may be of particular interest for further studies since receptor kinases recognize pathogen effectors and their rapid evolution, reflected by large numbers of polymorphisms, may represent plants adaptation to a rapidly changing spectrum of pathogens in the arms race between them and their hosts (Karasov et al., 2014).

The large number of SNPs that we identified will be instrumental for the development of linkage and high-throughput association mapping approaches and for the expansion of our previous diversity studies in Lathyrus (Almeida et al., 2014; Vaz Patto and Rubiales, 2014).

104 Grass pea response to rust ______

Our results provide an overview of gene expression profiles of contrasting L. sativus genotypes inoculated with rust, offering a valuable set of sequence data for candidate rust resistant gene discovery.

3.6. Conclusions

Our transcriptome analysis provided comprehensive insight into the molecular mechanisms underlying prehaustorial rust resistance in L. sativus.

The differences in resistance between the two L. sativus genotypes investigated appear to be mainly due to the activation of the SA pathway and several pathogenesis related genes, including the ones regulated by MLO. The fastest-evolving pathways differentiating between the two genotypes are the general RNA’s regulation of transcription, followed by the Ubiquitin-26S proteasome system and having also as most mutated receptor-based signalling genes and biotic and abiotic stress related genes. The detected polymorphic SNPs will allow the development of new gene-based molecular tools. Altogether, 51 genes were identified as potential resistance genes, prioritizing them as specific targets for future functional studies on grass pea/rust interactions. Besides a plethora of pathogenesis-related host genes, 4,558 transcripts, including putative effectors, were also identified for the rust fungus U. pisi. As a consequence of the newly developed wider array of genetic and genomic resources, future work will focus on high throughput mapping of the genetic basis of disease resistance in L. sativus and eventual comparative mapping with other legume species,

105 Chapter 3 ______contributing all to an improved exploitation of this under used highly potential legume species.

3.7. Acknowledgments

NFA carried out the inoculations and sample processing for sequencing and RT-qPCR, designed the RT-qPCR primers, performed the data analysis, developed the EST-SSRs primers and drafted the manuscript. NFA and STL participated in the results generation on RT-qPCR and data analysis. STL performed the EST- SSR genotyping. NK and BR performed the RNA sequencing and bioinformatics data processing. PW and DR revised the manuscript critically. MCVP coordinated the study and participated to the drafting and revision of the manuscript. NFA would like to thank the CRF- INIA, Madrid, Spain, for supplying the genotypes.

3.8. References

Almeida, N.F., Leitão, S.T., Caminero, C., Torres, A.M., Rubiales, D., and Vaz Patto, M.C. (2014). Transferability of molecular markers from major legumes to Lathyrus spp. for their application in mapping and diversity studies. Molecular Biology Reports 41, 269-283. Bari, R., and Jones, J.G. (2009). Role of plant hormones in plant defence responses. Plant Molecular Biology 69, 473-488. Barilli, E., Moral, A., Sillero, J.C., and Rubiales, D. (2012). Clarification on rust species potentially infecting pea (Pisum sativum L.) crop and host range of Uromyces pisi (Pers.) Wint. Crop Protection 37, 65-70. Barilli, E., Satovic, Z., Rubiales, D., and Torres, A.M. (2010a). Mapping of quantitative trait loci controlling partial resistance against rust incited by Uromyces pisi (Pers.) Wint. in a Pisum fulvum L. intraspecific cross. Euphytica 175, 151-159. Barilli, E., Satovic, Z., Sillero, J.C., Rubiales, D., and Torres, A.M. (2011). Phylogenetic analysis of Uromyces species infecting

106 Grass pea response to rust ______

grain and forage legumes by sequence analysis of nuclear ribosomal internal transcribed spacer region. Journal of Phytopathology 159, 137-145. Barilli, E., Sillero, J.C., and Rubiales, D. (2010b). Induction of systemic acquired resistance in pea against rust (Uromyces pisi) by exogenous application of biotic and abiotic inducers. Journal of Phytopathology 158, 30-34. Bennett, M.D., and Leitch, I.J. (2012). "Angiosperm DNA C-values database". (release 8.0, Dec. 2012) http://data.kew.org/cvalues/). Campbell, C.G. (1997). Promoting the conservation and use of underutilized and neglected crops. Grass pea. Lathyrus sativus L. Rome, Italy: International Plant Genetic Resources Institute. Cao, F.Y., Yoshioka, K., and Desveaux, D. (2011). The roles of ABA in plant-pathogen interactions. Journal of Plant Research 124, 489-499. Chen, X., Coram, T., Huang, X., Wang, M., and Dolezal, A. (2013). Understanding molecular mechanisms of durable and non- durable resistance to stripe rust in wheat using a transcriptomics approach. Current Genomics 14, 111-126. Chowdhury, M.A., and Slinkard, A.E. (1997). Natural outcrossing in grasspea. Journal of Heredity 88, 154-156. Drummond, A.J., Ashton, B., Buxton, S., Cheung, M., Cooper, A., Duran, C., Field, M., Heled, J., Kearse, M., Markowitz, S., Moir, R., Stones-Havas, S., Sturrock, S., Thierer, T., and Wilson, A. (2011). "Geneious v5.4, Available from http://www.geneious.com). Duke, J. (1981). Handbook of Legumes of World Economic Importance. New York, U.S.A.: Plennun Press. Durieu, P., and Ochatt, S.J. (2000). Efficient intergeneric fusion of pea (Pisum sativum L.) and grass pea (Lathyrus sativus L.) protoplasts. Journal of Experimental Botany 51, 1237-1242. Edgar, R.C. (2010). Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460-2461. Faleiro, F., Ragagnin, V.A., Moreira, M.A., and De Barros, E.G. (2004). Use of molecular markers to accelerate the breeding of common bean lines resistant to rust and anthracnose. Euphytica 138, 213-218. Fernandez, D., Tisserant, E., Talhinhas, P., Azinheira, H., Vieira, A.N.A., Petitot, A.-S., Loureiro, A., Poulain, J., Da Silva, C., Silva, M.D.O.C., and Duplessis, S. (2012). 454- pyrosequencing of Coffea arabica leaves infected by the rust fungus Hemileia vastatrix reveals in planta-expressed

107 Chapter 3 ______

pathogen-secreted proteins and plant functions in a late compatible plant–rust interaction. Molecular Plant Pathology 13, 17-37. Garcia, A., Calvo, E.S., De Souza Kiihl, R.A., Harada, A., Hiromoto, D.M., and Vieira, L.G. (2008). Molecular mapping of soybean rust (Phakopsora pachyrhizi) resistance genes: discovery of a novel locus and alleles. Theoretical and Applied Genetics 117, 545-553. Ghosh, M. (2006). Antifungal properties of haem peroxidase from Acorus calamus. Annals of Botany 98, 1145-1153. Giraldo, M.C., and Valent, B. (2013). Filamentous plant pathogen effectors in action. Nature Reviews Microbiology 11, 800-814. Glazebrook, J. (2005). Contrasting mechanisms of defense against biotrophic and necrotrophic pathogens. Annual Review of Phytopathology 43, 205-227. Guo, Z., Bonos, S., Meyer, W., Day, P., and Belanger, F. (2003). Transgenic creeping bentgrass with delayed dollar spot symptoms. Molecular Breeding 11, 95-101. Gutiérrez-Marcos, J.F., Vaquero, F., Sáenz De Miera, L.E., and Vences, F.J. (2006). High genetic diversity in a world-wide collection of Lathyrus sativus L. revealed by isozymatic analysis. Plant Genetic Resources 4, 159-171. Gutterson, N., and Reuber, T.L. (2004). Regulation of disease resistance pathways by AP2/ERF transcription factors. Current Opinion in Plant Biology 7, 465-471. Hacquard, S., Veneault-Fourrey, C., Delaruelle, C., Frey, P., Martin, F., and Duplessis, S. (2011). Validation of Melampsora larici- populina reference genes for in planta RT-quantitative PCR expression profiling during time-course infection of poplar leaves. Physiological and Molecular Plant Pathology 75, 106- 112. Hillocks, R., and Maruthi, M. (2012). Grass pea (Lathyrus sativus): Is there a case for further crop improvement? Euphytica 186, 647-654. Huang, X., and Madan, A. (1999). CAP3: A DNA sequence assembly program. Genome Research 9, 868-877. Hyten, D.L., Hartman, G.L., Nelson, R.L., Frederick, R.D., Concibido, V.C., Narvel, J.M., and Cregan, P.B. (2007). Map location of the Rpp1 locus that confers resistance to soybean rust in soybean. Crop Science 47, 837-840. Jørgensen, I.H. (1992). Discovery, characterization and exploitation of Mlo powdery mildew resistance in barley. Euphytica 63, 141-152.

108 Grass pea response to rust ______

Karasov, T.L., Horton, M.W., and Bergelson, J. (2014). Genomic variability as a driver of plant–pathogen coevolution? Current Opinion in Plant Biology 18, 24-30. Kemen, E., Gardiner, A., Schultz-Larsen, T., Kemen, A.C., Balmuth, A.L., Robert-Seilaniantz, A., Bailey, K., Holub, E., Studholme, D.J., Maclean, D., and Jones, J.D. (2011). Gene gain and loss during evolution of obligate parasitism in the white rust pathogen of Arabidopsis thaliana. PLoS Biol 9, e1001094. Kemen, E., Kemen, A., Ehlers, A., Voegele, R., and Mendgen, K. (2013). A novel structural effector from rust fungi is capable of fibril formation. The Plant Journal 75, 767-780. Kemen, E., Kemen, A.C., Rafiqi, M., Hempel, U., Mendgen, K., Hahn, M., and Voegele, R.T. (2005). Identification of a protein from rust fungi transferred from haustoria into infected plant cells. Molecular Plant-Microbe Interactions 18, 1130-1139. Kim, D.S., and Hwang, B.K. (2012). The pepper MLO gene, CaMLO2, is involved in the susceptibility cell-death response and bacterial and oomycete proliferation. The Plant Journal 72, 843-855. Kim, M.C., Panstruga, R., Elliott, C., Muller, J., Devoto, A., Yoon, H.W., Park, H.C., Cho, M.J., and Schulze-Lefert, P. (2002). Calmodulin interacts with MLO protein to regulate defence against mildew in barley. Nature 416, 447-451. Link, T.I., Lang, P., Scheffler, B.E., Duke, M.V., Graham, M.A., Cooper, B., Tucker, M.L., Van De Mortel, M., Voegele, R.T., Mendgen, K., Baum, T.J., and Whitham, S.A. (2014). The haustorial transcriptomes of Uromyces appendiculatus and Phakopsora pachyrhizi and their candidate effector families. Molecular Plant Pathology 15, 379-393. Liu, C.J., Huhman, D., Sumner, L.W., and Dixon, R.A. (2003). Regiospecific hydroxylation of isoflavones by cytochrome p450 81E enzymes from Medicago truncatula. The Plant Journal 36, 471-484. Lohse, M., Nagel, A., Herter, T., May, P., Schroda, M., Zrenner, R., Tohge, T., Fernie, A.R., Stitt, M., and Usadel, B. (2014). Mercator: a fast and simple web server for genome scale functional annotation of plant sequence data. Plant Cell and Environment 37, 1250-1258. Madrid, E., Rubiales, D., Moral, A., Moreno, M.T., Millán, T., Gil, J., and Rubio, J. (2008). Mechanism and molecular markers associated with rust resistance in a chickpea interspecific cross (Cicer arietinum × Cicer reticulatum). European Journal of Plant Pathology 121, 43-53.

109 Chapter 3 ______

Marino, D., Peeters, N., and Rivas, S. (2012). Ubiquitination during plant immune signaling. Plant Physiology 160, 15-27. Marioni, J.C., Mason, C.E., Mane, S.M., Stephens, M., and Gilad, Y. (2008). RNA-seq: an assessment of technical reproducibility and comparison with gene expression arrays. Genome Research 18, 1509-1517. Marrs, K.A. (1996). The functions and regulation of glutathione S- transferases in plants. Annu Rev Plant Physiol Plant Mol Biol 47, 127-158. Martín-Sanz, A., Palomo, J., Pérez De La Vega, M., and Caminero, C. (2012). Characterization of Pseudomonas syringae pv. syringae isolates associated with bacterial blight in Lathyrus spp. and sources of resistance. European Journal of Plant Pathology 134, 205-216. Mayer, C. (2010) Phobos 3.3.11, http://www.rub.de/spezzoo/cm/ cm_phobos.htm. Mcdonald, B., and Linde, C. (2002). The population genetics of plant pathogens and breeding strategies for durable resistance. Euphytica 124, 163-180. Mcgrann, G.R.D., Stavrinides, A., Russell, J., Corbitt, M.M., Booth, A., Chartrain, L., Thomas, W.T.B., and Brown, J.K.M. (2014). A trade off between mlo resistance to powdery mildew and increased susceptibility of barley to a newly important disease, Ramularia leaf spot. Journal of Experimental Botany. Medzhitov, R., and Janeway Jr, C.A. (1997). Innate immunity: the virtues of a nonclonal system of recognition. Cell 91, 295-298. Melotto, M., Underwood, W., Koczan, J., Nomura, K., and He, S.Y. (2006). Plant stomata function in innate immunity against bacterial invasion. Cell 126, 969-980. Metzker, M.L. (2010). Sequencing technologies - the next generation. Nature Reviews Genetics 11, 31-46. Miklas, P., Kelly, J., Beebe, S., and Blair, M. (2006). Common bean breeding for resistance against biotic and abiotic stresses: From classical to MAS breeding. Euphytica 147, 105-131. Morant, A.V., Bjarnholt, N., Kragh, M.E., Kjærgaard, C.H., Jørgensen, K., Paquette, S.M., Piotrowski, M., Imberty, A., Olsen, C.E., Møller, B.L., and Bak, S. (2008). The β-glucosidases responsible for bioactivation of hydroxynitrile glucosides in Lotus japonicus. Plant Physiology 147, 1072-1091. Nicaise, V., Roux, M., and Zipfel, C. (2009). Recent advances in PAMP-triggered immunity against bacteria: pattern recognition receptors watch over and raise the alarm. Plant Physiology 150, 1638-1647.

110 Grass pea response to rust ______

O’connell, R.J., and Panstruga, R. (2006). Tête à tête inside a plant cell: establishing compatibility between plants and biotrophic fungi and oomycetes. New Phytologist 171, 699-718. Pieterse, C.M.J., Van Der Does, D., Zamioudis, C., Leon-Reyes, A., and Van Wees, S.C.M. (2012). Hormonal modulation of plant immunity. Annual Review of Cell and Developmental Biology, Vol 28 28, 489-521. Pretsch, K., Kemen, A., Kemen, E., Geiger, M., Mendgen, K., and Voegele, R. (2013). The rust transferred proteins-a new family of effector proteins exhibiting protease inhibitor function. Molecular Plant Pathology 14, 96-107. Qin, X., and Zeevaart, J.a.D. (1999). The 9-cis-epoxycarotenoid cleavage reaction is the key regulatory step of abscisic acid biosynthesis in water-stressed bean. Proceedings of the National Academy of Sciences of the United States of America 96, 15354-15361. Rahman, M.M., Kumar, J., Rahman, M.A., and Ali Afzal, M. (1995). Natural outcrossing in Lathyrus sativus L. Indian Journal of Genetics & Plant Breeding 55, 204-207. Rai, R., Singh, A.K., Singh, B.D., Joshi, A.K., Chand, R., and Srivastava, C.P. (2011). Molecular mapping for resistance to pea rust caused by Uromyces fabae (Pers.) de-Bary. Theoretical and Applied Genetics 123, 803-813. Rojas-Molina, M.D., Rubiales, D., Prats, E., and Sillero, J.C. (2007). Effects of phenylpropanoid and energetic metabolism inhibition on faba bean resistance mechanisms to rust. Phytopathology 97, 60-65. Roth, A., Ding, J., Morin, R., Crisan, A., Ha, G., Giuliany, R., Bashashati, A., Hirst, M., Turashvili, G., Oloumi, A., Marra, M.A., Aparicio, S., and Shah, S.P. (2012). JointSNVMix: a probabilistic model for accurate detection of somatic mutations in normal/tumour paired next-generation sequencing data. Bioinformatics 28, 907-913. Rubiales, D., Castillejo, M.A., Madrid, E., Barilli, E., and Rispail, N. (2011). Legume breeding for rust resistance: lessons to learn from the model Medicago truncatula. Euphytica 180, 89-98. Rubiales, D., and Niks, R.E. (1995). Characterization of Lr34, a major gene conferring nonhypersensitive resistance to wheat leaf rust. Plant Disease 79, 1208-1212. Rubiales, D., Sillero, J.C., and Emeran, A.A. (2013). Response of vetches (Vicia spp.) to specialized forms of Uromyces vicia- fabae and to Uromyces pisi. Crop Protection 46, 38-43.

111 Chapter 3 ______

Schmidt, S.M., and Panstruga, R. (2011). Pathogenomics of fungal plant parasites: what have we learnt about pathogenesis? Current Opinion in Plant Biology 14, 392-399. Schuelke, M. (2000). An economic method for the fluorescent labeling of PCR fragments. Nature Biotechnology 18, 233-234. Selitrennikoff, C.P. (2001). Antifungal proteins. Applied and Environmental Microbiology 67, 2883-2894. Shimada, S., Takahashi, K., Sato, Y., and Sakuta, M. (2004). Dihydroflavonol 4-reductase cDNA from non-anthocyanin- producing species in the Caryophyllales. Plant Cell Physiology 45, 1290-1298. Sillero, J.C., Fondevilla, S., Davidson, J., Patto, M.C.V., Warkentin, T.D., Thomas, J., and Rubiales, D. (2006). Screening techniques and sources of resistance to rusts and mildews in grain legumes. Euphytica 147, 255-272. Sillero, J.C., Rojas-Molina, M.M., Avila, C.M., and Rubiales, D. (2012). Induction of systemic acquired resistance against rust, ascochyta blight and broomrape in faba bean by exogenous application of salicylic acid and benzothiadiazole. Crop Protection 34, 65-69. Sillero, J.C., and Rubiales, D. (2002). Histological characterization of resistance to Uromyces viciae-fabae in faba bean. Phytopathology 92, 294-299. Simon, S.A., Zhai, J., Nandety, R.S., Mccormick, K.P., Zeng, J., Mejia, D., and Meyers, B.C. (2009). Short-read sequencing technologies for transcriptional analyses. Annual Review of Plant Biology 60, 305-333. Skiba, B., Ford, R., and Pang, E.C.K. (2005). Construction of a cDNA library of Lathyrus sativus inoculated with Mycosphaerella pinodes and the expression of potential defence-related expressed sequence tags (ESTs). Physiological and Molecular Plant Pathology 66, 55-67. Smith, J.L., De Moraes, C.M., and Mescher, M.C. (2009). Jasmonate- and salicylate-mediated plant defense responses to insect herbivores, pathogens and parasitic plants. Pest Management Science 65, 497-503. Son, G.H., Wan, J., Kim, H.J., Nguyen, X.C., Chung, W.S., Hong, J.C., and Stacey, G. (2011). Ethylene-Responsive Element- Binding Factor 5, ERF5, is involved in chitin-induced innate immunity response. Molecular Plant-Microbe Interactions 25, 48-60. Stergiopoulos, I., and De Wit, P.J.G.M. (2009). Fungal effector proteins. Annual Review of Phytopathology 47, 233-263.

112 Grass pea response to rust ______

Stintzi, A., Heitz, T., Prasad, V., Wiedemann-Merdinoglu, S., Kauffmann, S., Geoffroy, P., Legrand, M., and Fritig, B. (1993). Plant ‘pathogenesis-related’ proteins and their role in defense against pathogens. Biochimie 75, 687-706. Strickler, S.R., Bombarely, A., and Mueller, L.A. (2012). Designing a transcriptome next-generation sequencing project for a nonmodel plant species. American Journal of Botany 99, 257- 266. Thimm, O., Blasing, O., Gibon, Y., Nagel, A., Meyer, S., Kruger, P., Selbig, J., Muller, L.A., Rhee, S.Y., and Stitt, M. (2004). MAPMAN: a user-driven tool to display genomics data sets onto diagrams of metabolic pathways and other biological processes. The Plant Journal 37, 914-939. Tisserant, E., Kohler, A., Dozolme-Seddas, P., Balestrini, R., Benabdellah, K., Colard, A., Croll, D., Da Silva, C., Gomez, S.K., Koul, R., Ferrol, N., Fiorilli, V., Formey, D., Franken, P., Helber, N., Hijri, M., Lanfranco, L., Lindquist, E., Liu, Y., Malbreil, M., Morin, E., Poulain, J., Shapiro, H., Van Tuinen, D., Waschke, A., Azcón-Aguilar, C., Bécard, G., Bonfante, P., Harrison, M.J., Küster, H., Lammers, P., Paszkowski, U., Requena, N., Rensing, S.A., Roux, C., Sanders, I.R., Shachar- Hill, Y., Tuskan, G., Young, J.P.W., Gianinazzi-Pearson, V., and Martin, F. (2012). The transcriptome of the arbuscular mycorrhizal fungus Glomus intraradices (DAOM 197198) reveals functional tradeoffs in an obligate symbiont. New Phytologist 193, 755-769. Untergasser, A., Cutcutache, I., Koressaar, T., Ye, J., Faircloth, B.C., Remm, M., and Rozen, S.G. (2012). Primer3-new capabilities and interfaces. Nucleic Acids Research 40, e115. Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A., and Speleman, F. (2002). Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biology 3, research0034-research0034.0011. Varshney, R.K., Graner, A., and Sorrells, M.E. (2005). Genic microsatellite markers in plants: features and applications. Trends in Biotechnology 23, 48-55. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2006a). Characterization of resistance to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus sativus. Plant Breeding 125, 308-310. Vaz Patto, M.C., and Rubiales, D. (2009). Identification and characterization of partial resistance to rust in a germplasm collection of Lathyrus sativus L. Plant Breeding 128, 495-500.

113 Chapter 3 ______

Vaz Patto, M.C., and Rubiales, D. (2014). Lathyrus diversity: available resources with relevance to crop improvement – L. sativus and L. cicera as case studies. Annals of Botany 113, 895-908. Vaz Patto, M.C., Skiba, B., Pang, E.C.K., Ochatt, S.J., Lambein, F., and Rubiales, D. (2006b). Lathyrus improvement for resistance against biotic and abiotic stresses: From classical breeding to marker assisted selection. Euphytica 147, 133- 147. Vleeshouwers, V.G., Raffaele, S., Vossen, J.H., Champouret, N., Oliva, R., Segretin, M.E., Rietman, H., Cano, L.M., Lokossou, A., Kessel, G., Pel, M.A., and Kamoun, S. (2011). Understanding and exploiting late blight resistance in the age of effectors. Annual Review of Phytopathology 49, 507-531. Wang, L., Feng, Z., Wang, X., and Zhang, X. (2010). DEGseq: an R package for identifying differentially expressed genes from RNA-seq data. Bioinformatics 26, 136-138. Westermann, A.J., Gorski, S.A., and Vogel, J. (2012). Dual RNA-seq of pathogen and host. Nature Reviews Microbiology 10, 618- 630. Zheng, Z., Nonomura, T., Appiano, M., Pavan, S., Matsuda, Y., Toyoda, H., Wolters, A.-M.A., Visser, R.G.F., and Bai, Y. (2013). Loss of function in Mlo orthologs reduces susceptibility of pepper and tomato to powdery mildew disease caused by Leveillula taurica. PLoS ONE 8, e70723. Zurbriggen, M.D., Carrillo, N., and Hajirezaei, M.R. (2010). ROS signaling in the hypersensitive response: when, where and what for? Plant Signal Behav 5, 393-396.

114 Grass pea response to ascochyta blight ______

Chapter 4 Lathyrus sativus transcriptome resistance response to Ascochyta lathyri investigated by deepSuperSAGE analysis

The work presented in this chapter was mostly performed by Nuno Felipe Almeida (see Acknowledgements section), and corresponds to the following publication:

Almeida, N.F., Krezdorn, N., Rotter, B., Winter, P., Rubiales, D., and Vaz Patto, M.C. (2015). Lathyrus sativus transcriptome resistance response to Ascochyta lathyri investigated by deepSuperSAGE analysis. Frontiers in Plant Science 6, 178.

115

Grass pea response to ascochyta blight ______

4.1. Abstract

Lathyrus sativus (grass pea) is a temperate grain legume crop with a great potential for expansion in dry areas or zones that are becoming more drought-prone. It is also recognized as a potential source of resistance to several important diseases in legumes, such as ascochyta blight. Nevertheless, the lack of detailed genomic and/or transcriptomic information hampers further exploitation of grass pea resistance-related genes in precision breeding. To elucidate the pathways differentially regulated during ascochyta-grass pea interaction and to identify resistance candidate genes, we compared the early response of the leaf gene expression profile of a resistant L. sativus genotype to Ascochyta lathyri infection with a non- inoculated control sample from the same genotype employing deepSuperSAGE. This analysis generated 14.387 UniTags of which 95.7% mapped to a reference grass pea/rust interaction transcriptome. From the total mapped UniTags, 738 were significantly differentially expressed between control and inoculated leaves. The results indicate that several gene classes acting in different phases of the plant/pathogen interaction are involved in the L. sativus response to A. lathyri infection. Most notably a clear up-regulation of defence- related genes involved in and/or regulated by the ethylene pathway was observed. There was also evidence of alterations in cell wall metabolism indicated by overexpression of cellulose synthase and lignin biosynthesis genes. This first genome-wide overview of the gene expression profile of the L. sativus response to ascochyta infection delivered a valuable set of candidate resistance genes for future use in precision breeding.

117 Chapter 4 ______

4.2. Introduction

Lathyrus sativus (grass pea) is a diploid species (2n = 14; genome size of approx. 8.2 Gbp (Bennett and Leitch, 2012)) with a great potential for expansion in dry areas or zones that are becoming more drought-prone (Hillocks and Maruthi, 2012). This species has been also recognized as a potential source of resistance to several important diseases in legumes (Vaz Patto and Rubiales, 2014).

Ascochyta blights are among the most important plant diseases worldwide (Rubiales and Fondevilla, 2012). Among the legume species, ascochyta blights are incited by different pathogens. For example, ascochytoses are caused by Ascochyta rabiei (teleomorph Didymella rabiei) in chickpea, A. fabae (teleomorph D. fabae) in faba bean and A. lentis (teleomorph D. lentis) in lentil (Tivoli et al., 2006). Ascochyta blight in pea (Pisum sativum) is caused by a fungal complex formed by A. pisi, A. pinodes (teleomorph Didymella pinodes (syn. Mycosphaerella pinodes)) and Phoma medicaginis var. pinodella (Jones, 1927). Of these, D. pinodes is the most frequent and damaging (Tivoli and Banniza, 2007).

Lathyrus spp. (L. sativus, L. cicera, L. ochrus and L. clymenum) however, are significantly more resistant to D. pinodes than field pea cultivars (Gurung et al., 2002). A detailed analysis of quantitative resistance of L. sativus to ascochyta blight, caused by D. pinodes, suggested that resistance in L. sativus may be controlled by two independently segregating genes, operating in a complementary epistatic manner (Skiba et al., 2004b). In another study, Skiba et al. (2004a) developed a grass pea linkage map and used it to locate two quantitative trait loci (QTL), explaining 12% and 9% of the observed variation in resistance to D. pinodes. Nevertheless, no candidate genes were identified at that time for these resistance QTLs,

118 Grass pea response to ascochyta blight ______hampering their use in precision breeding. In an attempt to identify defence-related candidate genes involved in D. pinodes resistance in L. sativus, the expression of 29 potentially defence-related ESTs was compared between L. sativus resistant and susceptible lines (Skiba et al., 2005). These ESTs were selected from a previously developed cDNA library of L. sativus stem and leaf tissue challenged with D. pinodes. From these, sixteen ESTs were considered eventually important for conferring stem resistance to ascochyta blight in L. sativus. In addition, the marker developed from one of them, EST LS0574 (Cf-9 resistance gene cluster), was significantly linked to one of the previously identified resistance QTLs. However this study was necessarily limited to the small number of initially selected EST sequences.

deepSuperSAGE (Matsumura et al., 2012) is the combination of SuperSAGE (Matsumura et al., 2003) with high-throughput sequencing technologies, allowing genome-wide and quantitative gene expression profiling. Two recent studies applied this technique for the identification of genes involved in resistance to ascochyta blight in pea (Fondevilla et al., 2014) and faba bean (Madrid et al., 2013).

In the present study we employed deepSuperSAGE to obtain a genome-wide overview of the response of the transcriptome of a resistant L. sativus genotype to A. lathyri infection in comparison to a non-inoculated control. Thereby we aimed at elucidation of signaling pathways responding to A. lathyri infection and identification of candidate genes associated with resistance to ascochyta blight in grass pea as first step towards the development of effective strategies for legume resistance breeding against this pathogen.

119 Chapter 4 ______

4.3. Materials and methods

4.3.1. Plant material and inoculation

Lathyrus sativus genotype BGE015746, previously characterized by our team as resistant to A. lathyri (isolate “Asc.8”), not developing macroscopic disease symptoms (pers. comm.), was used for the experiments. Isolate “Asc.8” belongs to the fungal collection of the Institute for Sustainable Agriculture-CSIC (Córdoba, Spain) while the L. sativus genotype BGE015746 was kindly provided by the Plant Genetic Resources Centre (CRF-INIA), Madrid, Spain. Fifteen-days old seedlings, grown in plastic pots containing 250 cm3 of 1:1 sand-peat mixture in a controlled growth chamber (20 ± 2 ºC with a 12 h light photoperiod), were inoculated with the monoconidial A. lathyri isolate “Asc.8”, collected in Zafra, Spain. Three individual plants were used for each treatment (inoculated/control). Spore suspension for inoculation was prepared at a concentration of 5x105 spores per millilitre and sprayed onto the plants’ aerial parts as described by Fondevilla et al. (2014). Inoculated and control plants were then kept in the dark for 24 h at 20 ºC and with 100% relative humidity in order to promote spore germination and were then transferred to the initial growth chamber conditions. Resistance was confirmed by the absence of disease symptoms 15 days after inoculation (d.a.i.), while other Lathyrus spp. genotypes presented diverse levels of infection, ranging up to 60% of leaf area covered by lesions (pers. comm.).

4.3.2. RNA extraction and deepSuperSAGE library construction

Leaves from one plant per treatment were harvested at 2 h time intervals during the first 24 h after inoculation (h.a.i.). A total of

120 Grass pea response to ascochyta blight ______twelve leaf samples per plant (one per each 2 h time point) were immediately frozen in liquid nitrogen after harvest and stored at -80 ºC. Total RNA was isolated from each sample separately, using the GeneJet Plant purification kit (Thermo Scientific, Vilnius, Lithuania) according to the manufacture’s protocols. Isolated RNA was subsequently treated with Turbo DNase I (Ambion, Austin, TX, USA), and quantified by NanoDrop (Thermo Scientific, Passau, Germany). 100 µg-samples of individual plant RNA from each time point were then pooled in two bulks, a control and an inoculated pool. RNA integrity was controlled by electrophoresis on a 2% agarose gel (Lonza, Rockland, USA) with SYBRSafe (Invitrogen, Eugene, USA) staining and visualized using a GEL-DOC 1000 System (Bio-Rad, Hercules, USA). deepSuperSAGE libraries from the two pools of control and inoculated RNAs were generated at GenXPro GmbH as described by Zawada et al. (2011). High-throughput DNA sequencing was performed on an Illumina Genome Analyser IIx using the Chrysalis 36 cycles v 4.0 sequencing kit. The multiplexed sequencing run consists of thirty-eight sequencing cycles on a single lane.

4.3.3. Data analysis and annotation

The sequence reads obtained by Illumina sequencing from each of the two pooled samples were processed with GenXPro’s in- house analysis pipeline. Briefly, libraries were sorted according to their respective index, followed by elimination of PCR-derived tags identified by TrueQuant technology. The sequences representing distinct deepSuperSAGE tags were quantified. These unique sequences (UniTags) were subsequently annotated against various databases via BLAST (Altschul et al., 1990). A multi-step BLAST procedure was used to annotate the UniTag reads to ensure an

121 Chapter 4 ______unambiguous assignment to their corresponding transcript and to eliminate any remaining adaptor sequences. Reference datasets were generated by own de-novo-assembly (Almeida et al., 2014) and downloaded from the publicly accessible Fabaceae databases using the nucleotide database from the National Center for Biotechnology Information (NCBI). UniTag reads were successively aligned against these reference datasets in the following order: (1) 26 bp de-novo- assembly dataset with a minimum BLAST-score of 42; (2) UniTags which did not attain the specified BLAST score in the previous step were aligned against the complete NCBI dataset with the same required BLAST score of 42 or above. For each library, UniTag read numbers were normalized to a million sequenced reads in total (tags per million; TPM) to allow the comparison between the two (control/inoculated) libraries. P-values for the UniTags were calculated using a perl module (http://search.cpan.org/~scottzed/Bio- SAGE-Comparison-1.00/) (Velculescu et al., 1995; Audic and Claverie, 1997; Saha et al., 2002). The fold changes were calculated as the log2 ratio of the normalized values between the two libraries.

4.3.4. Quantitative RT-PCR assays

For the quantitative RT-PCR assay, RNA samples from the different time points were pooled into two composite samples per plant, one control and one inoculated, in equimolar amounts. One μg of total RNA from each of these six composite samples (three plants/ two treatments) was reverse transcribed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA), according to manufacturer’s instructions. For all studied genes, the product of each of these reactions was analysed in technical duplicates, in a total of six technical replicates per treatment

122 Grass pea response to ascochyta blight ______

(inoculated/control). Analysed genes were selected by their level of expression and tag count from the deepSuperSAGE analysis. The chosen UniTags differed between inoculated and control samples by log2 ratios ranging from -1.73 to 3.37, with UniTag counts ranging from 1 to 558. Primers were designed using the Primer3 software (Untergasser et al., 2012) (Table 4.1), and RT-qPCR reactions performed with an iQ™5 Real-Time PCR Detection System (Bio-Rad, Munich, Germany). Data analysis was performed using the Genex software package (MultiD, Goteborg, Sweden), by the geNorm software (Vandesompele et al., 2002).

4.3.5. UniTag assignment to functional categories

In order to classify the L. sativus UniTags into functional categories, the Mercator pipeline for automated sequence annotation (Lohse et al., 2014), available at http://mapman.gabipd.org/web/ guest/app/mercator, was employed. The mapping file was created using only significantly (-1 ≤ log2 fold change ≥ 1; p-value < 0.05) up- and down-regulated UniTags and accessing the following, manually curated databases: Arabidopsis TAIR proteins (release 10), SwissProt/UniProt Plant Proteins (PPAP), TIGR5 rice proteins (ORYZA), Clusters of orthologous eukaryotic genes database (KOG), Conserved domain database (CCD) and InterPro scan (IPR). The Mercator mapping file was then employed for analysis by the MapMan software (Thimm et al., 2004), available at http://mapman.gabipd.org/web/guest/mapman.

123 124 ______Chapter 4

Table 4.1. Contig information and primer sequences for RT-qPCR.

Contig BLAST hit Forward primer 5´3´ Reverse primer 5´3´

1-acyl-sn-glycerol-3-phosphate a7162;272 CATGCTTTGCTTCGGTGGAC CTTGGACCGCCATTGCAATC acyltransferase 1, chloroplastic-like a600;793 40S ribosomal protein S24 GCGGACAAGGCAGTCACTAT GGCCTTTGAGACATTAGCCCT

a574;578 40S ribosomal protein S29 ATGGACTCATGTGCTGCAGG AAACCTAACCTTGGCTGGCC ABC transporter A family a11456;203 TGCATCCATCATGGTGACGG TGCTGCCCAGTTTCACTGTT member 1-like Abhydrolase domain containing a14590;204 CCCGACAGTGAATCCCTTCC ACAGACAGCAGTGCCGAAAT protein a6507;507 β-tubulin TGCCTAGGATCAGCAGCACA TCAGTGTCCCTGAGCTCACT

a5354;161 histone H3 ACGCTCGCCTCTAATACGC GCAGCTGAGTCGTACCTTGT

a833;622 L-allo-threonine aldolase AGTCACGGAATCACCAAATCCC ATCGTCTCGTGGCTTGTGG

a6396;394 Malic enzyme TTGGCTACGCATCTTCCTCG GCTTCTGTTCACCTATAGTTGCGG oxygen-evolving enhancer a156;1828 GTACTTCTCTGCTTCTGAGGGAC CCAAGCCTAAGGACCAGAAACA protein 3 precursor a8658;358 primary amine oxidase-like GGGCCTTTCAAAGCTTGGC TGTTCCTCCAAGCCCAAGTG RNA polymerase II C-terminal a18319;129 AATCTCGCGATCCACGTCAC TGGCTTGTGGAACGAATGAGG domain phosphatase-like protein a1255;508 unknown AGTGCGGGTATGGAATCACG TGGGACACCAGATGAATGGC

a10868;260 Villin-4 GTCAGCTCCCGGCAGTTTAG AAAGTTTCCCGGGAGCAGTC

Grass pea response to ascochyta blight ______

4.4. Results

4.4.1. SuperSAGE library characterization

A total of 399,648 deepSuperSAGE 26bp-tags were obtained. Of these 205,691 tags were derived from L. sativus inoculated with A. lathyri and 193,957 tags from control plants. These tags corresponded to 14,386 unique sequences (UniTags) of which 13,773 (95.7%) were successfully annotated to the L. sativus reference dataset (Almeida et al., 2014).

When comparing inoculated versus control samples, 738 UniTags were differentially expressed (DE) (log2 fold ≥ 2 (up) or log2 fold ≤ -2 (down); p-value < 0.05). Of the differentially expressed UniTags, 625 (84.7%) were successfully annotated in public plant databases. 354 UniTags matched also to entries in fungal databases, but bit scores were always lower than the plant database hit, and therefore were considered UniTags of plant origin. From the 625 differentially expressed UniTags with BLAST hit, 382 (61.1%) were up-regulated while 243 (38.9%) were down-regulated. The full list of differentially expressed UniTags can be found in Additional file 4.1.

4.4.2. SuperSAGE validation by quantitative RT-qPCR assay

From the geNorm software analysis the best housekeeping gene for the quantitative RT-qPCR validation was “β-tubulin” (transcript a6507;507). The expression levels of the remaining thirteen genes analysed by RT-qPCR to validate the RNA-Seq results are present in Table 4.2. A good correlation (R=0.8) was observed between the log2 fold changes measured by deepSuperSAGE and RT-qPCR for the genes tested (Figure 4.1).

125 Chapter 4 ______

Figure 4.1: Relative expression levels correlation between RNA-Seq and RT-qPCR. Pearson’s correlation coefficient (R) between relative expression levels is shown below the trend line.

4.4.3. Annotation of differentially expressed genes in the resistant L. sativus genotype after A. lathyri infection

Functional annotation of the UniTags via Mercator and MapMan, grouped 625 UniTags (382 up- and 243 down-regulated) into 25 main functional categories. Most represented categories from up-regulated UniTags were “protein metabolism” (11.6% up- and 8.1% down-regulated), “RNA metabolism” (9.4% up- and 4.7% down- regulated), “miscellaneous” (5.7% up- and 3.4% down-regulated), “signaling” (4.7% up- and 3.4% down-regulated) and “cell metabolism” (4.2% up- and 2.7% down-regulated) (Figure 4.2).

Potential candidate genes assigned to stress-related functional categories are listed in Table 4.3.

126 ______Table 4.2. Log2 fold expression results for deepSuperSAGE and RT-qPCR experiments.

BGE015746 inoculated/ Contig BLAST hit Control Inoculated inoculated/ control Control Inoculated control tags per tags per deepSuperSAGE tag counts tag counts qPCR million million (log2) (log2) 1-acyl-sn-glycerol-3-phosphate a7162;272 acyltransferase 1, chloroplastic- 1 5,1558 11 53,4783 3.375 1.81 like a600;793 40S ribosomal protein S24 76 391,839 57 277,115 -0.500 0.15

a574;578 40S ribosomal protein S29 171 881,639 128 622,293 -0.503 0.21 ABC transporter A family a11456;203 10 51,5578 23 111,818 1.117 0.23 member 1-like Abhydrolase domain containing a14590;204 5 25,7789 4 19,4466 -0.407 -0.01 protein

a5354;161 histone H3 54 278,412 31 150,712 -0.885 -0.79 Grass pea response to ascochyta

a833;622 L-allo-threonine aldolase 12 61,8694 36 175,02 1.500 0.45

a6396;394 Malic enzyme 18 92,8041 41 199,328 1.103 0.77 oxygen-evolving enhancer a156;1828 558 2876,93 292 1419,61 -1.019 -1.02 protein 3 precursor a8658;358 primary amine oxidase-like 122 629,005 39 189,605 -1.730 -2.38 RNA polymerase II C-terminal a18319;129 domain phosphatase-like 31 159,829 26 126,403 -0.338 0.21 protein a1255;508 unknown 36 185,608 36 175,02 -0.085 -0.74

a10868;260 Villin-4 30 154,673 14 68,0633 -1.184 0.04 blight 127

Chapter 4 ______

Figure 4.2: Percentage of annotated up- and down-regulated L. sativus UniTags upon A. lathyri inoculation in several functional categories by MapMan.

4.5. Discussion

The present study provides the first comprehensive overview of gene expression of the L. sativus response to ascochyta infection. It delivered a valuable set of grass pea sequences for resistance candidate gene discovery and use in precision breeding for this species.

deepSuperSAGE analysis of an ascochyta blight resistant grass pea genotype, using control and inoculated plants, generated 14.387 UniTags. Of those, 95.7% mapped to a recently published reference grass pea/rust interaction transcriptome assembly (Almeida et al., 2014). From the total mapped UniTags, 738 were differentially expressed between control and inoculated conditions, 625 of which could be annotated in public plant databases.

128 Grass pea response to ascochyta blight ______

Table 4.3 - List of detected genes by functional category, expression values and putative function as described in Mercator. log2 Functional Up/Down- Gene Contig fold Putative function category regulated change stress total of 13 genes multidrug and toxin compound extrusion a10578;232 10.1 transport (MATE) efflux armadillo (ARM) repeat superfamily a11957;197 9.2 protein degradation protein acidic endochitinase precursor (E.C. a3844;425 2.7 antimicrobial activity 3.2.1.14) Up RESISTANCE TO P. SYRINGAE PV biotic a15229;117 2.7 pathogen recognition MACULICOLA 1 (RPM1) pathogenesis related - function PR-1-like protein a8364;304 2.5 unknown disease resistance-responsive (dirigent- pathogenesis related - function a19559;148 2.4 like protein) unknown

Down similar to a chitin-binding protein (PR-4) a4526;396 -2.7 antimicrobial activity

DNAJ heat shock protein a14646;184 9.2 protein folding

Up DNAJ heat shock protein a11774;196 2.9 protein folding

heat shock protein 101 family a22155;174 2.7 thermotolerance to chloroplasts abiotic S-adenosyl-L-methionine-dependent a76762;48 -10.0 methylation methyltransferases superfamily proteins S-adenosyl-L-methionine-dependent Down a15131;185 -9.3 methylation methyltransferases superfamily proteins negative regulation of damaged DNA binding protein 1A a7531;356 -2.4 photomorphogenesis secondary total of 8 genes metabolism flavonoids Up chalcone reductase a124260;32 9.6 flavonoid biosynthesis

violaxanthin de-epoxidase a1039;529 10.2 isoprenoid biosynthesis

tocopherol cyclase a708;558 9.8 isoprenoid biosynthesis Up isoprenoids beta-hydroxylase 1 a3019;480 9.2 isoprenoid biosynthesis

RAB geranylgeranyl transferase beta a18716;198 2.7 isoprenoid biosynthesis subunit 1 pyridoxal phosphate (PLP)-dependent Down a23319;167 -9.3 isoprenoid biosynthesis transferases superfamily protein

phenylalanine ammonia-lyase 1 (PAL1) a5118;361 3.1 lignin biosynthesis phenylpropanoids Up 4-coumarate-CoA ligase a9524;336 2.4 lignin biosynthesis cell wall total of 9 genes precursor Up UDP-sugar pyrophospharylase a18802;199 9.2 cell wall synthesis synthesis

Up IRREGULAR XYLEM 1 (IRX1) a12901;208 2.9 cell wall synthesis

cellulose synthase isomer (CESA3) a6154;437 -9.0 cell wall synthesis cellulose synthesis Down cellulose-synthase-like C5 (CSLC5) a69762;64 -9.0 cell wall synthesis

glycosylphosphatidylinositol-anchored a5236;405 -2.2 cell wall synthesis protein COBRA-like (COB) degradation Down β-xylosidase 1 (BXL1) a4868;387 -2.5 cell wall degradation

xyloglucan endotransglycosylase-related modification Down a2002;437 -2.9 cell wall modifications protein (XTR4)

Up SKU5 similar 9 (sks9) a34641;119 9.4 cell wall modifications pectin*esterases plant invertase/pectin methylesterase Down a7441;320 -2.4 cell wall modifications inhibitor superfamily

129 Chapter 4 ______

Table 4.3 (cont.) log2 Functional Up/Down- Gene Contig fold Putative function category regulated change hormone total of 6 genes metabolism basic helix-loop-helix (bHLH) DNA- a246012;14 9.8 induced by ethylene binding superfamily protein 1-aminocyclopropane-1-carboxylate Up a11244;194 9.4 ethylene biosynthesis synthase (ACC) ethylene calmodulin-binding transcription activator a7309;224 8.9 induced by ethylene protein with CG-1 and Ankyrin domains inhibitor of ethylene Down RING E3 ligase, XBAT32 a25116;120 -2.9 biosynthesis salicylic acid Up UDP-glucosyltransferase 74F1 a25008;62 9.2 salicylic acid biosynthesis

plasma membrane protein KOBITO abscisic acid signal abscisic acid Up a7591;247 2.7 (KOB1) transduction miscella- total of 4 genes neous glutathione S Up glutathione S-transferase a20761;129 9.6 detoxification transferases production of reactive oxygen peroxidases Down peroxidase superfamily protein a20130;174 -9.3 species

Up glucan endo-1,3-beta-glucosidase a243608;22 8.9 antimicrobial activity beta 1,3 glucan hydrolases glucan endo-1,3-beta-glucosidase 11- Down a34766;162 -3.1 antimicrobial activity like

Although differences may be observed between deepSuperSAGE and RT-qPCR results due to the presence of different transcript isoforms from the same gene, or different genes from the same family that cannot be distinguished by the 26-bp tag of the 3’-untranslated region provided by deepSuperSAGE (Fondevilla et al., 2014), the validation of thirteen differentially expressed genes by RT-qPCR, using three biological replicates, provided a good correlation with deepSuperSAGE results. Interestingly, the most invariably expressed UniTag corresponded to a β-tubulin transcript. This transcript was also identified as the best normalization gene in a previous RNA-Seq study, where this genotype (BGE015746) was inoculated with Uromyces pisi (Almeida et al., 2014).

The functional interpretation of differential gene expression patterns provided evidence for the involvement of genes assigned to several functional categories in different phases of the plant/pathogen interaction. As listed in Table 4.3, the most significant stress-related

130 Grass pea response to ascochyta blight ______responses of the resistant genotype, however, were probably the clear-cut up-regulation of the ethylene signaling pathway represented by genes involved in ethylene synthesis and down-regulation of inhibitors of ethylene synthesis and the up-regulation of ethylene- induced genes. Another prominent response concerned alterations in the cell wall metabolism, as indicated by the up-regulation of cellulose synthase genes and genes related to lignin biosynthesis. Pathogenesis-related functions induced by ascochyta infection are discussed below.

4.5.1. Pathogen perception

The first step in plant defence response is pathogen detection by pattern recognition receptors (PRR) as part of the innate immune system. This pathogen perception will trigger signaling events that activate a broad array of downstream defensive measures in the plant (Nicaise et al., 2009). In this study we identified several differentially expressed receptor kinases (up- and down-regulated) containing leucine-rich repeats (LRRs), that are key players in the regulation of diverse biological processes such as development, hormone perception and/or plant defence (Torii, 2004). We also identified an up-regulated receptor kinase with a thaumatin-like domain (a36033;97, log2 fold = 9.4). Thaumatin is a pathogenesis related (PR) protein described as increasing the permeability of fungal membranes by pore-forming mechanisms and therefore restraining fungal growth or even killing it (Selitrennikoff, 2001). Several thaumatin-like proteins have been shown to increase resistance in potato (Acharya et al., 2013), rice (Datta et al., 1999), wheat (Anand et al., 2003) and grapevine (Jayasankar et al., 2003) to diverse fungal pathogens. Several transcription factors were also induced upon

131 Chapter 4 ______pathogen recognition. One “WRKY DNA-binding protein 4” (a8940;191, log2 fold = 8.9) was identified in our study as up- regulated after inoculation. WRKY transcription factors are induced after the recognition by intracellular receptors of pathogen virulence molecules (effectors). After its induction, WRKY transcription factors can positively or negatively regulate various aspects of pathogen- associated molecular pattern (PAMP)-triggered immunity (PTI) and effector triggered immunity (ETI) (review by Eulgem (2005)). Also related to ETI, we found an up-regulated transcript with homology to Arabidopsis “RESISTANCE TO P. SYRINGAE PV MACULICOLA 1 (RPM1)”, known to confer resistance to Pseudomonas syringae strains containing the avirulence genes avrB and avrRpm1 (Bisgrove et al., 1994). In the incompatible interaction in the model plant, RIN4 (RPM1 interacting protein 4) interacts with RPM1, to prevent its activation. Reduction of RIN4 expression enhances resistance to P. syringae and to the oomycete Paranospora parasitica. Therefore RIN4 is considered a negative regulator of basal plant defences that is activated by P. syringae´s avrB and avrRpm1 (Mackey et al., 2002). Assuming a similar function of the RPM1-homolog in grass pea-ascochyta interaction this gene could be a resistance-steering candidate gene. It would be further interesting to know whether up- regulation of the RPM1-homolog is part of a broad defence response, or if it is activated by a specific Ascochyta spp. effector that the grass pea’s RPM1 is able to recognize.

4.5.2. Hormone signaling

It is generally accepted that biotrophic pathogens usually trigger the salicylic acid (SA) pathway, while necrotrophic pathogens activate jasmonic acid (JA) and the ethylene (ET) pathways

132 Grass pea response to ascochyta blight ______

(Glazebrook, 2005; Bari and Jones, 2009). The nature of the initial phases of Ascochyta spp. infection in grass pea is still not completely understood. Normally considered as necrotroph, there is evidence, at least for some Ascochyta spp., for an early biotrophic phase spanning from the penetration of the epidermis of the plant until the initial colonization of the mesophyll (Tivoli and Banniza, 2007).

Our data, however, demonstrate that the ethylene pathway may have a major role in resistance of at least our grass pea accession to A. lathyri, in line with the necrotrophic nature of the interaction. For example, the “1-aminocyclopropane-1-carboxylate synthase (ACC)” gene involved in ET biosynthesis and other two genes described by Mercator (Lohse et al., 2014) as being induced by ethylene (“Calmodulin-binding transcription activator with CG-1 and Ankyrin domains” and “basic helix-loop-helix (bHLH) DNA- binding superfamily protein”) were significantly up-regulated upon infection. The transcript is homologous to the “Calmodulin-binding transcription activator with CG-1 and Ankyrin domains” previously identified as similar to “Calmodulin-binding protein/ER66 protein” from tomato (Skiba et al., 2005). It seems that in grass pea either different transcript isoforms or a gene family exists, since in Skiba et al. (2005), sixteen defence-related ESTs were identified with a greater or/and earlier expression in stems of resistant L. sativus genotypes compared with susceptible ones upon ascochyta blight inoculation. In our study from those only “Calmodulin-binding transcription activator with CG-1 and Ankyrin domains” was up-regulated whereas three other SuperTags with similar annotation were not differentially expressed. The incongruence between our results and that of Skiba et al. (2005) may be explained by the different mechanism of resistance, since the L. sativus genotype used by Skiba et al. (2005),

133 Chapter 4 ______

ATC 80878, is partially resistant, and the genotype used in our study, BGE015746, displays complete resistance. Furthermore, the pathogen isolates used in both studies were also different, since the ATC 80878 genotype was inoculated with a mixture of three highly aggressive (on several P. sativum genotypes) M. pinodes isolates (WAL3, T16 and 4.9) whereas our ascochyta inoculum was a monoconidial A. lathyri isolate.

Additionally in our study, “RING E3 ligase, XBAT32”, an ubiquitin described as negative regulator of ET biosynthesis in Arabidopsis during plant growth, development and salt stress (Prasad and Stone, 2010), was down-regulated again stressing the importance of ET for resistance in our L. sativus genotype. ET pathway induction was also observed by microarray and deepSuperSAGE analyses during the response of a resistant pea genotype to ascochyta blight infection (Fondevilla et al., 2011; Fondevilla et al., 2014). Thus, up-regulation of ET signaling may be a general response of temperate legumes to ascochyta blight infection.

Although the ET pathway was the only hormone pathway clearly up-regulated, other genes involved in hormone signaling were also up-regulated. These included “UDP-glycosyltransferase 74 F1 (UGT74F1)”, and “phenylalanine ammonia-laser 1” (PAL1), both involved in SA biosynthesis in Arabidopsis (Mauch-Mani and Slusarenko, 1996).

4.5.3. Cell wall fortification

Ascochyta lathyri penetrates the host’s epidermal cells via an as yet imperfectly described biotrophic or necrotrophic phase to reach the mesophyll. However, it is known that during pathogen penetration,

134 Grass pea response to ascochyta blight ______the plant’s cell wall is not just a static physical barrier. The perception of cell wall degradation by the pathogen can activate local plant responses that trigger repair and fortification mechanisms via expression of different genes as e.g. the cell wall synthesis precursor, “UDP-sugar pyrophosphorylase” (Gibeaut, 2000) or the cellulose synthase “IRREGULAR XYLEM 1 (IRX1)” genes both involved in cell wall synthesis (Taylor et al., 2000). Both were up-regulated in our grass pea genotype after A. lathyri inoculation. IRX1 was also up- regulated in the same genotype BGE015746 in response to the infection with rust (Almeida et al., 2014), suggesting that the induction of this cellulose synthase, and consequently cell wall strengthening, may play an important role in resistances of this grass pea genotype to diverse pathogens. Improving the cell wall lignin content is another common plant defence mechanism. In our study, inoculation elicited the expression of three UniTags representing genes implicated in cell lignification: “4-coumarate-CoA ligase”, involved in lignin biosynthesis (Lee et al., 1995), “disease resistance-responsive (dirigent-like)”, previously identified as improving lignin content at infection sites (Zhu et al., 2007) and a “phenylalanine ammonia-lyase 1 (PAL1)”, previously related to SA biosynthesis and also to the synthesis of lignin precursors (Mauch-Mani and Slusarenko, 1996).

However, there were also cell wall synthesis genes that were down-regulated. For example, three transcripts involved in cellulose biosynthesis (“glycosylphosphatidylinositol-anchored protein COBRA- like (COB)”, “cellulose synthase isomer (CESA3)” and “cellulose- synthase-like C5 (CSLC5)”) and two pectinesterase transcripts involved in cellulose biosynthesis and in cell wall modifications (Dai et al., 2011; Hansen et al., 2011; Liu et al., 2013), were down-regulated after inoculation. Though this is somehow unexpected in a resistant

135 Chapter 4 ______accession it may be explained by results from Arabidopsis where CESA3-deficient mutants reduced their cellulose synthesis, but instead activated lignin synthesis and defence responses through the jasmonate and the ethylene signaling pathways (Cano-Delgado et al., 2003; Hamann, 2012). These observations suggest that mechanisms monitoring cell wall integrity can activate lignification and defence responses. Therefore, cellulose biosynthesis may not only be involved in the first line of defence but also in signaling as an indirect defence mechanism. Histological analysis will allow clarifying this hypothesis in the future.

Additionally, “beta-xylosidase 1 (BXL1)” and “xyloglucan endotransglycosylase-related protein (XTR4)”, were found down- regulated after inoculation. BXL1 is involved in development of normal (non-infected) cell walls. BXL1 deficient Arabidopsis mutants showed alterations of cell wall composition and in plant development (Goujon et al., 2003). XTR4 belongs to the xyloglucan endotransglycosylase gene family, the so called endoxyloglucan transferases, that are involved in hemicellulose metabolism. Interestingly, XTR4 is down-regulated in Arabidopsis by the growth hormone auxin (Xu et al., 1996). Therefore, in grass pea these genes may be down-regulated under the mechanisms regulating cell wall thickening to restrict fungal penetration.

Taken together these results hint to a general reshuffling of cell wall components that exchanges certain cellulose types, restricts hemicelluloses and favours lignin as part of the resistance reaction of a resistant L. sativus genotype. To which extent these mechanisms contribute to resistance needs to be determined in populations segregating for resistance.

136 Grass pea response to ascochyta blight ______

4.5.4. Antimicrobial activity

Upon infection plants increase the production of antibacterial defence proteins to limit colonization by the pathogen (Consonni et al., 2009). After inoculation of grass pea with A. lathyri, “chitinase A (PR-3)” was up-regulated. Chitinases are involved in the inhibition of fungal hyphae growth in intercellular spaces as a defence response to fungal infection in several plant species (reviewed by Grover et al. (2012)). Additionally, a “GDSL lipase 1”, another antimicrobial compound that also functions as ET-dependent elicitor (Kwon et al., 2009), and a “pathogenesis-related protein (PR-1-like)” with antifungal properties (van Loon and van Strien, 1999) were up regulated. This PR-1-like transcript is similar to an EST sequence (DY396405) identified previously in the response of grass pea to M. pinodes (Skiba et al., 2005), but in that study it showed low to mid- level expression in leaf and stem tissue, with little difference between resistant and susceptible genotypes. PR-1-like genes were also up- regulated in the resistance response of our grass pea accession BGE015746 to rust infection (Almeida et al., 2014). Chitinases were also found up-regulated in the resistance response of pea to ascochyta blight infection (Fondevilla et al., 2014).

The phenylpropanoid secondary metabolite biosynthesis pathway is notorious for the production of antimicrobial compounds in plants. In our resistant genotype, inoculation elicited a “chalcone reductase” transcript coding for an enzyme that co-acts with chalcone synthase in the first step of flavonoid biosynthesis (Naoumkina et al., 2010). Interestingly, in a previous study in L. sativus a chalcone reductase EST was also up-regulated as a defence reaction after inoculation with M. pinodes (Skiba et al., 2005).

137 Chapter 4 ______

4.5.5. Reactive oxygen species

Reactive oxygen species (ROS) in plants are generated normally as by-products of oxidative phosphorylation and diverse biosynthetic pathways. Under non-stress conditions these potentially deleterious molecules are controlled by antioxidants. Under biotic or abiotic stress however, ROS production increases as part of the anti- microbial response. Their rapid accumulation of ROS creates an oxidative burst that may induce cell death and restricts the establishment of the pathogen in the plant (Apel and Hirt, 2004). In our study however, the lack of visual symptoms of a hypersensitive response or necrosis in the inoculated resistant grass pea, suggests that the over-production of ROS is not important for resistance in this plant/pathogen interaction. Moreover, our transcriptomic data reflects this aspect, since the only differentially expressed UniTag related to ROS regulation was a “peroxidase” which was down-regulated after inoculation.

4.5.6. Detoxification

During defence response, plants produce toxic compounds for defence and are themselves attacked by toxins secreted by the pathogen. To cope with toxins from the pathogen, plants developed several detoxification mechanisms. In our grass pea accession two UniTags related to detoxification were up-regulated upon A. lathyri infection, namely a “phytoene synthase”, a precursor in the carotenoids biosynthesis pathway and a “glutathione S-transferase (GST)”. Carotenoids are lipophilic antioxidants being able to detoxify various forms of ROS, playing an important role in both biotic and abiotic stress responses (Young, 1991; Ramel et al., 2012). GSTs

138 Grass pea response to ascochyta blight ______form a large family of enzymes that have diverse roles in detoxifying xenobiotics, antioxidant activity or ROS scavenging (Dalton et al., 2009). ROS scavengers are needed to maintain ROS activity levels below the oxidative damage threshold (Moller et al., 2007). GST was also found up-regulated upon inoculation in an ascochyta blight resistant pea genotype challenged with M. pinodes (Fondevilla et al., 2014), corroborating its important role in resistance.

4.6. Conclusions

Our deepSuperSAGE analysis provided deep insights into the molecular mechanisms underlying resistance to A. lathyri in L. sativus suggesting candidate genes and pathways potentially involved in ascochyta blight resistance in a particular, completely resistant genotype. Resistance reactions involved a wide range of reactions including changes in hormone signaling, biotic and abiotic stress reactions, cell wall metabolism and in the secondary metabolism that can now be further investigated. In particular, this study suggests a strong up-regulation of the ET pathway and of cell wall fortification upon inoculation with A. lathyri. In agreement with the macroscopic phenotypic observations 15 d.a.i., that gave no hint to the presence of an oxidative burst or hypersensitive response, the changes in transcripts related to ROS management were rather moderate. Thus, we conclude that the resistance of our L. sativus genotype BGE015746 to ascochyta is quantitative rather than qualitative, as it has been reported in other legume species such as pea (Carrillo et al., 2013), lentil (Tullu et al., 2006), faba bean (Rubiales et al., 2012) and chickpea (Hamwieh et al., 2013) and represents a potentially lasting source of resistance to ascochyta blight (Rubiales et al. 2015). To exploit this genotype for resistance breeding next steps will

139 Chapter 4 ______the identification of polymorphisms in the identified candidate resistance genes to facilitate resistance breeding by marker-assisted selection. On the other hand, we will use histological approaches to characterize in detail the type of resistance response and correlate it with the molecular mechanisms identified in this study. A deeper understanding of resistance mechanism and facilitated resistance breeding will help to harness grass pea for agronomy in dry areas or zones that are becoming more drought-prone due to global climate change in the future.

4.7. Acknowledgements

NFA carried out the inoculations and sample processing for sequencing and RT-qPCR, designed the RT-qPCR primers, performed the data analysis and drafted the manuscript. NK and BR performed the RNA sequencing and bioinformatics data processing. PW and DR revised the manuscript critically. MCVP coordinated the study and participated to the drafting and revision of the manuscript. NFA would like to thank the CRF-INIA, Madrid, Spain, for supplying the genotypes.

4.8. References Acharya, K., Pal, A.K., Gulati, A., Kumar, S., Singh, A.K., and Ahuja, P.S. (2013). Overexpression of Camellia sinensis thaumatin- like protein, CsTLP in potato confers enhanced resistance to Macrophomina phaseolina and Phytophthora infestans Infection. Molecular Biotechnology 54, 609-622. Almeida, N.F., Leitão, S.T., Krezdorn, N., Rotter, B., Winter, P., Rubiales, D., and Vaz Patto, M.C. (2014). Allelic diversity in the transcriptomes of contrasting rust-infected genotypes of Lathyrus sativus, a lasting resource for smart breeding. BMC Plant Biology 14, 376.

140 Grass pea response to ascochyta blight ______

Altschul, S.F., Gish, W., Miller, W., Myers, E.W., and Lipman, D.J. (1990). Basic local alignment search tool. Journal of Molecular Biology 215, 403-410. Anand, A., Zhou, T., Trick, H.N., Gill, B.S., Bockus, W.W., and Muthukrishnan, S. (2003). Greenhouse and field testing of transgenic wheat plants stably expressing genes for thaumatin- like protein, chitinase and glucanase against Fusarium graminearum. Journal of Experimental Botany 54, 1101-1111. Apel, K., and Hirt, H. (2004). Reactive oxygen species: metabolism, oxidative stress, and signal transduction. Annual Review of Plant Biology 55, 373-399. Audic, S., and Claverie, J.M. (1997). The significance of digital gene expression profiles. Genome Research 7, 986-995. Bari, R., and Jones, J.G. (2009). Role of plant hormones in plant defence responses. Plant Molecular Biology 69, 473-488. Bennett, M.D., and Leitch, I.J. (2012). "Angiosperm DNA C-values database". (release 8.0, Dec. 2012) http://data.kew.org/cvalues/). Bisgrove, S.R., Simonich, M.T., Smith, N.M., Sattler, A., and Innes, R.W. (1994). A disease resistance gene in Arabidopsis with specificity for two different pathogen avirulence genes. The Plant Cell 6, 927-933. Cano-Delgado, A., Penfield, S., Smith, C., Catley, M., and Bevan, M. (2003). Reduced cellulose synthesis invokes lignification and defence responses in Arabidopsis thaliana. The Plant Journal 34, 351-362. Carrillo, E., Rubiales, D., Perez-De-Luque, A., and Fondevilla, S. (2013). Characterization of mechanisms of resistance against Didymella pinodes in Pisum spp. European Journal of Plant Pathology 135, 761-769. Consonni, C., Humphry, M., and Panstruga, R. (2009). "Defence Responses in Plants," in Plant Relationships, ed. H. Deising. (Berlin Heidelberg, Germany: Springer), 363-385. Dai, X., You, C., Chen, G., Li, X., Zhang, Q., and Wu, C. (2011). OsBC1L4 encodes a COBRA-like protein that affects cellulose synthesis in rice. Plant Molecular Biology 75, 333-345. Dalton, D.A., Boniface, C., Turner, Z., Lindahl, A., Kim, H.J., Jelinek, L., Govindarajulu, M., Finger, R.E., and Taylor, C.G. (2009). Physiological roles of glutathione S-transferases in soybean root nodules. Plant Physiology 150, 521-530. Datta, K., Velazhahan, R., Oliva, N., Ona, I., Mew, T., Khush, G.S., Muthukrishnan, S., and Datta, S.K. (1999). Over-expression of the cloned rice thaumatin-like protein (PR-5) gene in transgenic rice plants enhances environmental friendly resistance to

141 Chapter 4 ______

Rhizoctonia solani causing sheath blight disease. Theoretical and Applied Genetics 98, 1138-1145. Eulgem, T. (2005). Regulation of the Arabidopsis defence transcriptome. Trends in Plant Science 10, 71-78. Fondevilla, S., Kuster, H., Krajinski, F., Cubero, J.I., and Rubiales, D. (2011). Identification of genes differentially expressed in a resistant reaction to Mycosphaerella pinodes in pea using microarray technology. BMC Genomics 12, 28. Fondevilla, S., Rotter, B., Krezdorn, N., Jüngling, R., Winter, P., and Rubiales, D. (2014). Identification of genes involved in resistance to Didymella pinodes in pea by deepSuperSAGE transcriptome profiling. Plant Molecular Biology Reporter 32, 258-269. Gibeaut, D.M. (2000). Nucleotide sugars and glycosyltransferases for synthesis of cell wall matrix polysaccharides. Plant Physiology and Biochemistry 38, 69-80. Glazebrook, J. (2005). Contrasting mechanisms of defence against biotrophic and necrotrophic pathogens. Annual Review of Phytopathology 43, 205-227. Goujon, T., Minic, Z., El Amrani, A., Lerouxel, O., Aletti, E., Lapierre, C., Joseleau, J.-P., and Jouanin, L. (2003). AtBXL1, a novel higher plant (Arabidopsis thaliana) putative beta-xylosidase gene, is involved in secondary cell wall metabolism and plant development. The Plant Journal 33, 677-690. Grover, A. (2012). Plant Chitinases: Genetic Diversity and Physiological Roles. Critical Reviews in Plant Sciences 31, 57- 73. Gurung, A.M., Pang, E.C.K., and Taylor, P.W.J. (2002). Examination of Pisum and Lathyrus species as sources of ascochyta blight resistance for field pea (Pisum sativum). Australasian Plant Pathology 31, 41-45. Hamann, T. (2012). Plant cell wall integrity maintenance as an essential component of biotic stress response mechanisms. Frontiers in Plant Science 3, 77. Hamwieh, A., Imtiaz, M., Hobson, K., and Ahmed Kemal, S. (2013). Genetic diversity of microsatellite alleles located at quantitative resistance loci for ascochyta blight resistance in a global collection of chickpea germplasm. Phytopathologia Mediterranea 52, 183-191. Hansen, K.D., Wu, Z., Irizarry, R.A., and Leek, J.T. (2011). Sequencing technology does not eliminate biological variability. Nature Biotechnology 29, 572-573.

142 Grass pea response to ascochyta blight ______

Hillocks, R., and Maruthi, M. (2012). Grass pea (Lathyrus sativus): Is there a case for further crop improvement? Euphytica 186, 647- 654. Jayasankar, S., Li, Z.J., and Gray, D.J. (2003). Constitutive expression of Vitis vinifera thaumatin-like protein after in vitro selection and its role in anthracnose resistance. Functional Plant Biology 30, 1105-1115. Jones, L.K. (1927). Studies of the nature and control of blight, leaf and pod spot, and footrot of peas caused by species of Ascochyta. New York State Agricultural Experiment Station Bulletin 547, 1- 45. Kwon, S.J., Jin, H.C., Lee, S., Nam, M.H., Chung, J.H., Kwon, S.I., Ryu, C.-M., and Park, O.K. (2009). GDSL lipase-like 1 regulates systemic resistance associated with ethylene signaling in Arabidopsis. The Plant Journal 58, 235-245. Lee, D., Ellard, M., Wanner, L., Davis, K., and Douglas, C. (1995). The Arabidopsis thaliana 4-coumarate:CoA ligase (4CL) gene: stress and developmentally regulated expression and nucleotide sequence of its cDNA. Plant Molecular Biology 28, 871-884. Liu, L., Shang-Guan, K., Zhang, B., Liu, X., Yan, M., Zhang, L., Shi, Y., Zhang, M., Qian, Q., Li, J., and Zhou, Y. (2013). Brittle Culm1, a COBRA-like protein, functions in cellulose assembly through binding cellulose microfibrils. PLoS Genetics 9, e1003704. Lohse, M., Nagel, A., Herter, T., May, P., Schroda, M., Zrenner, R., Tohge, T., Fernie, A.R., Stitt, M., and Usadel, B. (2014). Mercator: a fast and simple web server for genome scale functional annotation of plant sequence data. Plant Cell and Environment 37, 1250-1258. Mackey, D., Holt, B.F., 3rd, Wiig, A., and Dangl, J.L. (2002). RIN4 interacts with Pseudomonas syringae type III effector molecules and is required for RPM1-mediated resistance in Arabidopsis. Cell 108, 743-754. Madrid, E., Palomino, C., Plotner, A., Horres, R., Rotter, B., Winter, P., Krezdorn, N., and Torres, A.M. (2013). DeepSuperSage analysis of the Vicia faba transcriptome in response to Ascochyta fabae infection. Phytopathologia Mediterranea 52, 166-182. Matsumura, H., Molina, C., Krüger, D.H., Terauchi, R., and Kahl, G. (2012). "DeepSuperSAGE: High-Throughput Transcriptome Sequencing with Now- and Next-Generation Sequencing Technologies," in Tag-Based Next Generation Sequencing. Wiley-VCH Verlag GmbH & Co. KGaA), 1-21.

143 Chapter 4 ______

Matsumura, H., Reich, S., Ito, A., Saitoh, H., Kamoun, S., Winter, P., Kahl, G., Reuter, M., Kruger, D.H., and Terauchi, R. (2003). Gene expression analysis of plant host-pathogen interactions by SuperSAGE. Proceedings of the National Academy of Sciences of the United States of America 100, 15718-15723. Mauch-Mani, B., and Slusarenko, A.J. (1996). Production of salicylic acid precursors is a major function of phenylalanine ammonia- lyase in the resistance of Arabidopsis to Peronospora parasitica. The Plant Cell 8, 203-212. Moller, I.M., Jensen, P.E., and Hansson, A. (2007). Oxidative modifications to cellular components in plants. Annual Review of Plant Biology 58, 459-481. Naoumkina, M.A., Zhao, Q., Gallego-Giraldo, L., Dai, X., Zhao, P.X., and Dixon, R.A. (2010). Genome-wide analysis of phenylpropanoid defence pathways. Molecular Plant Pathology 11, 829-846. Nicaise, V., Roux, M., and Zipfel, C. (2009). Recent advances in PAMP-triggered immunity against bacteria: pattern recognition receptors watch over and raise the alarm. Plant Physiology 150, 1638-1647. Prasad, M.E., and Stone, S.L. (2010). Further analysis of XBAT32, an Arabidopsis RING E3 ligase, involved in ethylene biosynthesis. Plant Signaling & Behavior 5, 1425-1429. Ramel, F., Birtic, S., Ginies, C., Soubigou-Taconnat, L., Triantaphylidès, C., and Havaux, M. (2012). Carotenoid oxidation products are stress signals that mediate gene responses to singlet oxygen in plants. Proceedings of the National Academy of Sciences 109, 5535-5540. Rubiales, D., Avila, C.M., Sillero, J.C., Hybl, M., Narits, L., Sass, O., and Flores, F. (2012). Identification and multi-environment validation of resistance to Ascochyta fabae in faba bean (Vicia faba). Field Crops Research 126, 165-170. Rubiales, D., and Fondevilla, S. (2012). Future prospects for ascochyta blight resistance breeding in cool season food legumes. Frontiers in Plant Science 3. Rubiales, D., Fondevilla, S., Chen, W., Gentzbittel, L., Higgins, T.J.V., Castillejo, M.A., Singh, K.B., and Rispail, N. (2015). Achievements and challenges in legume breeding for pest and disease resistance. Critical Reviews in Plant Sciences 34, 195- 236. Saha, S., Sparks, A.B., Rago, C., Akmaev, V., Wang, C.J., Vogelstein, B., Kinzler, K.W., and Velculescu, V.E. (2002). Using the transcriptome to annotate the genome. Nature Biotechnology 20, 508-512.

144 Grass pea response to ascochyta blight ______

Selitrennikoff, C.P. (2001). Antifungal proteins. Applied and Environmental Microbiology 67, 2883-2894. Skiba, B., Ford, R., and Pang, E.C.K. (2004a). Construction of a linkage map based on a Lathyrus sativus backcross population and preliminary investigation of QTLs associated with resistance to ascochyta blight. Theoretical and Applied Genetics 109, 1726-1735. Skiba, B., Ford, R., and Pang, E.C.K. (2004b). Genetics of resistance to Mycosphaerella pinodes in Lathyrus sativus. Australian Journal of Agricultural Research 55, 953–960. Skiba, B., Ford, R., and Pang, E.C.K. (2005). Construction of a cDNA library of Lathyrus sativus inoculated with Mycosphaerella pinodes and the expression of potential defence-related expressed sequence tags (ESTs). Physiological and Molecular Plant Pathology 66, 55-67. Taylor, N.G., Laurie, S., and Turner, S.R. (2000). Multiple cellulose synthase catalytic subunits are required for cellulose synthesis in Arabidopsis. The Plant Cell 12, 2529-2539. Thimm, O., Blasing, O., Gibon, Y., Nagel, A., Meyer, S., Kruger, P., Selbig, J., Muller, L.A., Rhee, S.Y., and Stitt, M. (2004). MAPMAN: a user-driven tool to display genomics data sets onto diagrams of metabolic pathways and other biological processes. The Plant Journal 37, 914-939. Tivoli, B., and Banniza, S. (2007). Comparison of the epidemiology of ascochyta blights on grain legumes. European Journal of Plant Pathology 119, 59-76. Tivoli, B., Baranger, A., Avila, C.M., Banniza, S., Barbetti, M., Chen, W.D., Davidson, J., Lindeck, K., Kharrat, M., Rubiales, D., Sadiki, M., Sillero, J.C., Sweetingham, M., and Muehlbauer, F.J. (2006). Screening techniques and sources of resistance to foliar diseases caused by major necrotrophic fungi in grain legumes. Euphytica 147, 223-253. Torii, K.U. (2004). Leucine-rich repeat receptor kinases in plants: structure, function, and signal transduction pathways. International Review of Cytology 234, 1-46. Tullu, A., Tar'an, B., Breitkreutz, C., Banniza, S., Warkentin, T.D., Vandenberg, A., and Buchwaldt, L. (2006). A quantitative-trait locus for resistance to ascochyta blight [Ascochyta lentis] maps close to a gene for resistance to anthracnose [Colletotrichum truncatum] in lentil. Canadian Journal of Plant Pathology 28, 588-595. Untergasser, A., Cutcutache, I., Koressaar, T., Ye, J., Faircloth, B.C., Remm, M., and Rozen, S.G. (2012). Primer3-new capabilities and interfaces. Nucleic Acids Research 40, e115.

145 Chapter 4 ______

Van Loon, L.C., and Van Strien, E.A. (1999). The families of pathogenesis-related proteins, their activities, and comparative analysis of PR-1 type proteins. Physiological and Molecular Plant Pathology 55, 85-97. Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A., and Speleman, F. (2002). Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biology 3, research0034-research0034.0011. Vaz Patto, M.C., and Rubiales, D. (2014). Lathyrus diversity: available resources with relevance to crop improvement – L. sativus and L. cicera as case studies. Annals of Botany 113, 895-908. Velculescu, V.E., Zhang, L., Vogelstein, B., and Kinzler, K.W. (1995). Serial analysis of gene-expression. Science 270, 484-487. Xu, W., Campbell, P., Vargheese, A.K., and Braam, J. (1996). The Arabidopsis XET-related gene family: environmental and hormonal regulation of expression. The Plant Journal 9, 879- 889. Young, A.J. (1991). The photoprotective role of carotenoids in higher plants. Physiologia Plantarum 83, 702-708. Zawada, A.M., Rogacev, K.S., Rotter, B., Winter, P., Marell, R.R., Fliser, D., and Heine, G.H. (2011). SuperSAGE evidence for CD14++CD16+ monocytes as a third monocyte subset. Blood 118, e50-61. Zhu, L., Zhang, X., Tu, L., Zeng, F., Nie, Y., and Guo, X. (2007). Isolation and characterization of two novel dirigent-like genes highly induced in cotton (Gossypium barbadense and G. hirsutum) after infection by Verticillium dahliae. Journal of Plant Pathology 89, 41-45.

146 Chickling pea response to rust ______

Chapter 5

Differential expression and allelic diversity in Lathyrus cicera / rust infection, a comprehensive analysis

The work presented in this chapter was mostly performed by Nuno Felipe Almeida (see Acknowledgements section), and corresponds to the following manuscript in preparation: Almeida, N.F., Horres, R., Krezdorn, N., Leitão, S.T., Aznar- Fernandez, T., Rotter, B., Winter, P., Rubiales, D., and Vaz Patto, M.C. Differential expression and allelic diversity in Lathyrus cicera / rust infection, a comprehensive analysis.

147

Chickling pea response to rust ______

5.1. Abstract

Lathyrus cicera transcriptome, in response to rust infection, was analysed to unveil resistance mechanisms and develop novel molecular breeding tools, until now inexistent, for this robust legume species. RNA-Seq libraries were generated from control and rust- inoculated (Uromyces pisi) leaves from two L. cicera genotypes with contrasting resistance levels. A de novo assembly was performed in order to allow quantification and interpretation of transcripts differential expression and sequence polymorphisms. We analysed the profile of 111,024 transcripts upon inoculation with U. pisi. Functional annotation grouped 62.6% of the contigs present in plant databases (41.9% of total) in 33 main functional categories, being ‘protein’, ‘RNA’, ‘signalling’, ‘transport’ and ‘stress’ the most represented. Most differentially expressed transcripts upon inoculation in partially resistant and susceptible genotypes were involved in signalling, cell wall metabolism and synthesis of secondary metabolites. Several polymorphic EST-SSR and SNP markers between the two L. cicera genotypes were developed. Also allele-specific expression was detected and validated through specific dual labelled probes RT-qPCR assays. This study represents the first efforts for genomic precision breeding in L. cicera, providing a large new set of molecular markers and potential candidate resistance genes to rust infection.

5.2. Introduction

Lathyrus cicera L., known as chickling pea, is an annual legume belonging to the tribe Fabeae (Kenicer et al., 2005; Schaefer et al., 2012), and mainly grown as stock feed, both as fodder and grain (Hanbury et al., 1999). L. cicera can adapt well to harsh environments,

149 Chapter 5 ______being resistant to drought, water lodging and to several important legume biotic constrains. Among these we find resistance sources to rust (Vaz Patto et al., 2009), powdery mildew (Vaz Patto et al., 2007), bacterial blight (Martín-Sanz et al., 2012) and crenate broomrape (Fernández-Aparicio et al., 2009). In this way L. cicera is a good alternative for cropping systems in marginal lands and can function also as a source of resistance genes to related species such as pea (Vaz Patto et al., 2006).

Limited genetic resources exist for this plant species, hampering its potential fully exploitation in legume breeding. In the NCBI database, accessed on January 2015, we could only find 4 internal transcribed spacers (ITS), 1 antifungal protein DNA sequence, 1 convicilin gene sequence, 26 sequences from chloroplast regions and 4 protein amino-acid sequences. Although also in a reduced number, molecular tools from related species have proven to be useful in this legume species. 91 intron-targeted amplified polymorphic (ITAP), 14 expressed sequence tag Simple Sequence Repeats (EST- SSR), 10 genomic microsatellite (gSSR), 6 resistance genes analogs (RGA) and 7 disease resistance (DR) markers previously developed for other legume species (Medicago truncatula Gaertn., Pisum sativum L., Lens culinaris Medik., Lupinus spp. and Vicia faba L.) were successfully cross-amplified in L. cicera (Almeida et al., 2014a).

Rusts are among the most important legume diseases, being the Uromyces genus the most significant. Uromyces pisi (Pers.) Wint. is suggested to be the principal agent causing pea rust (Barilli et al., 2009), being also capable of causing infection in Lathyrus species (Vaz Patto et al., 2009; Vaz Patto and Rubiales, 2009) and in Vicia and Lens (Barilli et al., 2012; Rubiales et al., 2013).

150 Chickling pea response to rust ______

Lathyrus sativus is the genetically nearest cultivated species of L. cicera, with evidences suggesting that it even derivatives from L. cicera (Hopf, 1986). Consequently, these two species might share many physiological/genetic mechanisms in response to stress.

Previous studies demonstrated that both species show a compatible reaction to U. pisi. Also, similar defence mechanisms were present despite the different recorded resistance levels, being L. sativus generally more resistant than L. cicera accessions (Vaz Patto et al., 2009; Vaz Patto and Rubiales, 2009). Recently, a work unveiling L. sativus transcriptome in response to inoculation with U. pisi was published (Almeida et al., 2014b). The development of new tools will facilitate a comparative analysis between those two species resistance mechanisms.

Allele-specific gene expression, the differential expression of alleles, has been described in yeasts (Brem et al., 2002), mammals (Cowles et al., 2002; Yan et al., 2002) and plants (Guo et al., 2004; Zhang and Borevitz, 2009). Differential expression in allele variants may contribute to different phenotypes, such as resistance responses. Therefore the development of molecular tools that allow the detection and quantification of allele-specific expression will improve the understanding of the mechanisms regulating plant resistance to pathogens. In addition, allele-specific assays would be also be useful for mapping of related expression quantitative trait loci (eQTLs).

Understanding the genetic control of disease resistance in plants is important to develop efficient molecular breeding tools. These tools will be fundamental in the development of modern resistant cultivars but also in the introgression of their interesting resistance traits into related species.

151 Chapter 5 ______

The aims of this study were: 1) To detect potential genes involved in L. cicera resistance to rust infection. For this, L. cicera differentially expressed genes upon inoculation were functionally characterized and their expression profiles compared; 2) To develop novel molecular markers to be employed in future L. cicera mapping and diversity studies, especially EST-SSR and SNP (Single Nucleotide Polymorphism); 3) To examine the differential expression of allelic variants after inoculation, based on the SNP information within candidate genes, and develop appropriate assays for future eQTLs analysis associated with rust resistance in L. cicera.

5.3. Material and methods

5.3.1. Plant material, inoculation and DNA and RNA isolation

The two L. cicera genotypes, BGE008277 and BGE023542, used in the present work were kindly provided by the Plant Genetic Resources Centre (CRF-INIA), Madrid, Spain. Seeds were multiplied in insect proof cages. Previous evaluation of resistance levels against U. pisi inoculation had demonstrated that BGE008277 is susceptible to rust, whereas BGE023542 displays partial resistance (Vaz Patto et al., 2009). Upon infection, both genotypes present well-formed pustules, with no associated chlorosis or necrosis. However, clear differences exist in the percentage of leaf area covered by the fungus (disease severity; DS). While the partially resistant BGE023542 has a DS=36%, the susceptible BGE008277 has a DS=80%.

The U. pisi monosporic isolate UpCo-01, from the fungal collection of the Institute for Sustainable Agriculture-CSIC (Córdoba, Spain), was used for inoculation. Inoculum was multiplied on plants of the susceptible P. sativum cv. Messire before use.

152 Chickling pea response to rust ______

Two-week-old L. cicera seedlings were inoculated by dusting all the plants at the same time with 2 mg of spores per plant, diluted in pure talk (1:10), with the help of a small manual dusting device, in a complete random experiment. Twenty-four plants per each L. cicera genotype and treatment, inoculated and control (non-inoculated), were then incubated for 24 h at 20 ºC, in complete darkness, and 100% relative humidity, then transferred to a growth chamber and kept at 20 ± 2 ºC under 14h light (150 μmol m−2 s−1) and 10h dark.

RNA, to be used in the RNA-Seq experiment, RT-qPCR validation and SNP validation, was extracted from inoculated and non- inoculated fresh leaves, from each of the 24 individual plants/genotype/treatment, collected 37 hours after inoculation, immediately frozen in liquid nitrogen and stored at -80 °C. RNA was isolated using the GeneJET Plant RNA Purification Mini Kit (Thermo Scientific, Vilnius, Lithuania), according to the manufacturer’s instructions. Isolated RNA was treated with Turbo DNase I (Ambion, Austin, TX, USA), and RNA quantification was carried out using the NanoDrop device (Thermo Scientific, Passau, Germany).

For the EST-SSR validation, DNA from frozen young leaves, from the two L. cicera genotypes, was extracted using a modified CTAB protocol developed by Torres et al. (1993).

5.3.2. RNA sequencing and transcript quantification

For each of the 4 combinations, genotype and treatment (BGE008277 control, BGE008277 inoculated, BGE023542 control and BGE023542 inoculated) total RNA from 24 plants was pooled in equal amounts for sequencing. Five RNA-Seq libraries (one for each genotype, each one with the two treatments, and one reference

153 Chapter 5 ______assembly, including all genotypes and treatments) were generated by GenXPro GmbH, Germany, using a proprietary protocol. In short, for each library, mRNA was captured from 20 µg of total RNA using Oligo dT(25) beads (Dynabeads; life Technologies). The purified mRNA was randomly fragmented in a Zn2+ solution to obtain approximately 250 bp long RNA fragments. cDNA was synthesized by reverse transcription starting from 6(N) random hexamer oligonucleotides, followed by second strand synthesis. Barcoded Y-adapters were ligated to the cDNA and the library was amplified with 10 cycles of PCR. The libraries were sequenced on an Illumina Hiseq2000 machine. After Illumina paired-end sequencing, raw sequence reads were passed through quality filtering, thereby also removing sequencing adapter primers and cDNA synthesis primers. All high-quality reads were assembled using the Trinity RNA-Seq de novo assembly (Version: trinityrnaseq_r2011- 11-26). In order to minimize the redundancy CAP3 software (Huang and Madan, 1999) was also used with overlap length cutoff of 30 bp and overlap percent identity cutoff of 75%. Redundancy was tested using the clustering algorithm UCLUST ((Edgar, 2010), available at http://drive5.com/usearch/manual/uclust_algo.html). The resulting contigs were annotated via BLASTX to publically available plant databases (ftp://ftp.ncbi.nlm.nih.gov/blast/db/FASTA/nr.gz, nr, plants only). To identify potential fungal transcripts, an additional BLASTX to public fungal databases (http://www.ebi.ac.uk/uniprot, UniProtKB/Swiss-Prot and UniProtKB/TrEMBL) was performed. The sequenced reads were mapped with novoalign software (V2.07.14; http://www.novocraft.com/) to the own assembled contigs. RPKM (reads per kilobase per million) was calculated as the normalized transcript expression value (Marioni et al., 2008). The obtained counts were subsequently passed through DEGSeq to calculate the

154 Chickling pea response to rust ______differential gene expression (R package version 1.16.0) (Wang et al., 2010).

5.3.3. Contig annotation and data analysis

In order to classify the obtained contigs into functional categories, the Mercator pipeline for automated sequence annotation (Lohse et al., 2014), available at http://mapman.gabipd.org/ web/guest/app/mercator), was used. The mapping file was created with information from the following manually curated databases: Arabidopsis TAIR proteins (release 10), SwissProt/UniProt Plant Proteins (PPAP), TIGR5 rice proteins (ORYZA), Clusters of orthologous eukaryotic genes database (KOG), Conserved domain database (CDD) and InterPro scan (IPR). The Mercator mapping file was then employed for pathway analysis by the MapMan software (Thimm et al., 2004), available at http://mapman.gabipd.org/web/ guest/mapman).

Differentially expressed contigs were identified by comparing their expression in leaves of the partially resistant genotype BGE023542, control vs. inoculated, and of the susceptible genotype BGE008277, control vs. inoculated, using DEGseq (Wang et al., 2010). In cases where a particular transcript had the same profile in both genotypes, the total transcript count, before and after inoculation, was compared, allowing the identification of basal genotypic differences between the two genotypes.

155 Chapter 5 ______

5.3.4. RNA-Seq validation by quantitative RT-PCR assay

To validate the RNA-Seq results, expression levels of a set of 10 selected genes were analysed by RT-qPCR. Genes were selected by their level of expression and transcript count, in order to represent a broad range of expression profiles. Further, the number of their transcripts differed between inoculated and control samples by log2 ratios ranging from -2.19 to 3.31. Their read count numbers were generally higher than 100, with three exceptions. Contig a20510;122, ‘Histone H2A.2’, with 28 counts in BGE008277 inoculated and 88 counts in BGE023542 inoculated sample, and contig a6507;507 ‘β- tubulin’, and contig a77720;50 ‘γ-tubulin’, with 76 an 73 counts in the susceptible inoculated line, respectively (Additional file 5.1).

1 μg of total RNA from each of three randomly chosen plants per genotype, per treatment (inoculated/control), was reverse transcribed in duplicates, using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA) following manufacturer’s instructions. Three independent reverse-transcription reactions (RT) were performed for each cDNA sample in a total of nine samples per genotype, per treatment.

For all genes studied, the product of each of these reactions was analysed in technical replicates, in a total of six technical replicates per treatment. RT-qPCR reactions were performed with an iQ™5 Real- Time PCR Detection System (Bio-Rad, Munich, Germany). Primers were designed using the Primer3 software (Untergasser et al., 2012). Primer sequences can be found in Additional file 5.1. For data analysis, the Genex software package (MultiD, Goteborg, Sweden), using the NormFinder software (Andersen et al., 2004) was employed.

156 Chickling pea response to rust ______

5.3.5. SNP detection

SNPs were discovered between the two L. cicera genotypes using the software JointSNVMix (Roth et al., 2012). The mappings from the transcriptome analysis were also analysed by Joint-SNV-Mix and the output was furthermore processed by GenXPro’s in-house software to detect SNPs discriminating the variant alleles. SNP calling was performed taking into account only the inoculated samples. A minimum coverage of 15 reads in each genotype in the inoculated condition was needed to call a SNP.

5.3.6. Allele-specific expression analysis by dual labelled probe RT-qPCR assays

In order to analyse the differential expression profile between allelic variants, SNPs in genes of interest discriminating the genotypes BGE023542 and BGE008277 after inoculation were selected for the design of allele-specific RT-qPCR assays (see Additional file 5.2). The SNP-containing transcripts analysed included 11 transcripts that had the similar expression level between the genotypes (five of them are housekeeping genes), one transcript with higher expression level on the partially resistant genotype and four transcripts with higher expression level on the susceptible genotype. Altogether 2x 17 allele- specific dual labelled probe RT-qPCR assays with SNP specific mismatch primer were tested.

The primer design of the dual labelled probe RT-qPCR assays with introduced additional mismatch forward or reverse primer was made with short amplicon size of <90 bp in order to avoid background by flanking additional SNPs. See the documentation of the primer design in the Additional file 5.3. We made a one-step RT-qPCR with

157 Chapter 5 ______

40-60 ng total RNA template per reaction from BGE023542 and BGE008277 and the corresponding SNP alleles in separate tubes, so reactions for each allele with the same 5’6-Fam-3’TQ2 dual labelled probe and the different mismatch primer. All reactions were made in 12µl reaction volume with the One Step Prime Script™ RT-PCR Kit (Perfect Real Time) from TAKARA Bio Inc., Japan. The RT-qPCR regime consisted of a reverse transcription step of 5 min at 42ºC and a initial denaturation step of 10 s at 95ºC, followed by 40 cycles of 5 s at 95ºC (denaturation) and 30 s at 62ºC (annealing/elongation).

5.3.7. EST-SSR development and genotyping

EST-SSRs were searched in silico from the obtained transcriptomes employing Phobos (Mayer, 2010) plug-in for Geneious software (Drummond et al., 2011), and using as search parameters, perfect SSRs with a repeat unit length of two to six nucleotides. Length polymorphisms were manually identified by aligning SSR-containing contigs of one genotype against the whole library of the other genotype. EST-SSR development and genotyping procedures were conducted as described in Almeida et. al. (2014b).

5.4. Results

5.4.1. RNA-Seq transcriptomes of analysed L. cicera genotypes

RNA-Seq libraries from control and inoculated leafs from each L. cicera genotype were united, prior to assembly, to generate a comprehensive data set enabling the generation of contigs of maximum length. The susceptible genotype BGE008277 united library included 18,395,860 reads, which were assembled into 66,210 contigs, ranging in size from 150 to 8,664 bp, with a mean contig length of 537

158 Chickling pea response to rust ______bp. The united library from the partially resistant genotype BGE023542 comprised 30,320,831 reads which assembled in 64,382 contigs, with a size range of 150 to 9,694 bp and a mean contig length of 571 bp.

The reference assembly using both genotypes and treatments gather 145,985 contigs, ranging in size from 150 to 13,916 bp, with a mean contig length of 485 bp. The mapping and quantification of both genotypes’ libraries to the reference assembly allowed the analysis of their differential expression in response to U. pisi infection. 20,362 contigs were unique to the partially resistant and 12,114 contigs were unique to the susceptible genotype.

5.4.2. Differential gene expression in partially resistant and susceptible L. cicera genotypes to rust infection

Differentially expressed L. cicera contigs after rust inoculation were grouped by expression patterns based on up- or down-regulation (log2 ≥ 2 or log2 ≤ -2; respectively, q-value ≤ 0.05). Within each expression pattern group, comparisons were performed between genotypes. Expression patterns were grouped in eight response types, according to their up- or down-regulation, in susceptible and partially resistant genotypes, respectively, similarly to a previous work, where the transcriptomic response of two L. sativus genotypes with contrasting level of resistance were compared upon inoculation with U. pisi (Almeida et al., 2014b).

The number of differentially expressed contigs and description of each group is summarized in Table 5.1. Most representative groups are group F (contigs down regulated in both genotypes) and H (contigs down-regulated only in the partially resistant genotype) with 4,520 and 3,498 contigs respectively, followed by a group that includes 2,161

159 Chapter 5 ______contigs up-regulated upon infection in both partially resistant and susceptible genotypes (group A). A detailed list with all the identified contigs, their description and expression pattern groups can be found in Additional File 5.4.

Table 5.1 - Classification of contigs according to their differential expression in the susceptible and resistant genotype upon infection with U. pisi. Up regulated: (log2 >= 2; q- value ≤ 0.05); Down-regulated: (log2 ≤ -2; q-value ≤ 0.05); higher in Susceptible: (log2 fold change between all resistant and susceptible genotype contigs <= -2; q-value ≤ 0.05); higher in Resistant: (log2 fold change between all resistant and susceptible genotype contigs >= 2; q-value ≤ 0.05)

Expression pattern Feature # of contigs group

Up-regulated in Resistant A 2,161 Up-regulated in Susceptible Up-regulated in Resistant B 12 Up-regulated in Susceptible, higher in Susceptible Up-regulated in Resistant, higher in Resistant C 20 Up-regulated in Susceptible

D Up-regulated in Susceptible 1,715

E Up-regulated in Resistant 338

Down-regulated in Resistant F 4,520 Down-regulated in Susceptible

G Down-regulated in Susceptible 1,399

H Down-regulated in Resistant 3,498

Total 13,663

As depicted in Figure 5.1, from the 111,287 contigs that could be identified and quantified, 43,590 were shared among all libraries.

160 Chickling pea response to rust ______

Figure 5.1 - Venn diagram of the number of unique and shared contigs between the two genotypes and its expression. Partially resistant genotype: BGE023542, Susceptible genotype: BGE008277.

5.4.3. RNA-Seq validation by quantitative RT-PCR assay

To validate the RNA-Seq results, expression levels of a set of 10 selected genes were analysed by RT-qPCR. Genes were selected by their level of expression and transcript count, in order to represent a broad range of expression profiles. Furthermore, the number of their transcripts differed between inoculated and control samples by log2 ratios, ranging from -2.19 to 3.31. Their read count numbers were generally higher than 100, with three exceptions: Contig a20510;122, ‘Histone H2A.2’, with 28 counts in BGE008277 inoculated and 88 counts in BGE023542 inoculated sample, and contig a6507;507 ‘β- tubulin’, and contig a77720;50 ‘γ-tubulin’, with 76 an 73 counts in the susceptible inoculated line, respectively. The best reference gene assay for normalization, suggested by the NormFinder software, for

161 Chapter 5 ______both genotype samples, was a transcript coding for a ‘γ-tubulin’ (a77720;50). Results for the analysed transcripts can be found in Additional file 5.1.

5.4.4. Annotation of L. cicera contigs

From the 111,287 contigs detected in all libraries, 46,588 (41.9%) contigs were matched, via BLAST, to entries in plant databases and 622 (0.6%) contigs matched only to fungal databases, being present only in the inoculated libraries. In addition to the already described contigs, 688 (0.6%) other contigs absent in control samples had a higher bit-score in fungal databases than in plant databases and thus, most probably correspond also to U. pisi sequences.

Figure 5.2 - Number of contigs that could be BLASTed to different plant species

162 Chickling pea response to rust ______

As indicated in Figure 5.2, BLAST produced hits mainly to other legume species. Medicago truncatula (23,754; 50.99%), Cicer arietinum (11,177; 23.99%), Glycine max (3,559; 7.64%), P. sativum (1,224; 2.63%), Phaseolus vulgaris (800; 1.72%) and Lotus japonicus (300; 0.64%) were the best matching legume species. Vitis vinifera (1,128; 2.42%), Hordeum vulgare (307; 0.66%), Zea mays (216; 0.46%) and the model Arabidopsis thaliana (214; 0.46%) were the best matching non-legume species.

From the 49 accessions described in UniProtKB/Swiss-Prot and UniProtKB/TrEMBL as U. viciae-fabae, 12 were identified in our data, four of them absent in control samples and found exclusively in fungal databases, and eight had a higher bit-score in fungal databases than in plant databases (see list in Additional file 5.5. None of these 12 contigs were significantly differentially expressed between the two inoculated genotypes. For example, among the four contigs out of the 12 without a plant database hit, three were homologous to ‘invertase 1’, and the other to ‘rust transferred protein – Rtp1’.

Functional annotation of the all identified differentially expressed contigs via Mercator and MapMan, depicted in Figure 5.3, grouped them into 34 main functional categories, of which the categories ‘protein’ (11.5%), ‘RNA’ (8.7%), ‘signalling’ (6.4%), ‘transport’ (5.2%), ‘miscellaneous’ (4.7%), and ‘stress’ (3.9%) were the most enriched. A total of 37.4% differentially expressed (DE) contigs could not be assigned to any functional category.

When analysing in more detail the stress related functional category, we observed that transcripts involved in the several layers of defence against pathogens were assigned to different expression pattern groups (Thimm et al., 2004). In order to restrict the number of analysed contigs to the ones probably more directly related to disease

163 Chapter 5 ______

Figure 5.3 - Percentage of contigs assigned in each main functional category. resistance, the analysis was focused on contigs up-regulated in the partially resistant genotype, upon inoculation with U. pisi (group E). The complete set of transcripts and its expression profiles can be found in Additional file 5.4.

164 Chickling pea response to rust ______

In group E, several transcripts were identified as related with signaling and regulation of transcription of defence responses. Ten receptor kinases, one calcium receptor and a ‘WRKY family transcription factor’ were found up-regulated in the partially resistant genotype.

Plant hormones play a key role in numerous processes, including modulating the response to biotic stresses (Bari and Jones, 2009). In this particular expression pattern group E, only one transcript that may play a role in biotic stresses was identified, in the ‘hormone metabolism’ category, a ‘SAUR-like auxin-responsive protein family’ (a165751;26).

Plant cell wall is not just a static physical barrier. Related with cell wall degradation, a ‘glycosyl hydrolase superfamily protein’ (a77705;78), a ‘rhamnogalacturonate lyase family protein’ (a160424;29), a ‘polygalacturonase precursor (EC 3.2.1.15)’ (a385339;11) and a ‘pectinesterase-2 precursor (EC 3.1.1.11)’ (a13317;203) were identified in group E.

Upon infection, plants increase the production of defence proteins with antibacterial properties to limit the pathogen colonization (Consonni et al., 2009). In the subcategory ‘secondary metabolism’ four transcripts were identified: two transcripts encoding a 3-ketoacyl- CoA synthase family protein (KCS6) involved in the biosynthesis of very long chain fatty acids, in addition to a ‘hydroxycinnamoyl-CoA shikimate/quinate hydroxycinnamoyl transferase’ involved in the phenylpropanoid pathway; and a ‘DOWNY MILDEW RESISTANT 6 (DMR6)’ transcript, which was found to encode a 2-oxoglutarate (2OG)-Fe(II) oxygenase of unknown function, but that has been involved in defence responses to biotic stress (Van Damme et al., 2005; 2008).

165 Chapter 5 ______

Although no hypersensitive response was detected, a ‘peroxidase superfamily protein’, that generate hydrogen peroxide involved in oxidative stress (Passardi et al., 2005) was identified in group E. Furthermore, a ‘glucan endo-1,3-beta-glucosidase precursor (EC 3.2.1.39)’ was identified DE in group E.

5.4.5. SNP identification in resistance pathways

In the 43,590 transcripts present in both the susceptible and the partially resistant genotypes (Figure 5.1), 19,224 Single Nucleotide Polymorphisms (SNPs) in 5,152 transcripts were detected between the two genotypes in the inoculated samples. Among those, 811 contigs containing SNPs discriminating between their respective alleles were functionally annotated and identified. The number of SNPs in functional (MapMan) categories varied considerably. The categories ‘RNA regulation of transcription’ (6.15%) and ‘protein.degradation’ (4.97%) contained by far the most SNPs, followed by the protein-related categories ‘protein.postranslational modification’ (2.72%), ‘protein.synthesis’ (1.89%) and ‘protein.targeting’ (1.78%). Other categories among the most SNP-containing contigs were ‘signalling.receptor kinases’ (1.42%) and ‘hormone metabolism.auxin’ (1.30%) (Figure 5.4).Polymorphic contigs and their respective SNPs are listed in Additional file 5.6.

5.4.6. Allele-specific expression validation by dual probe assays

Seventeen allele-specific expression assays were analysed for SNP validation in the genotypes BGE023542 and BGE008277. These resulted for 5 sites in the confirmation of allele-specific expression for both alleles. Additionally the analysis of allele-specific expression was

166 Chickling pea response to rust ______nine times positive for one of the two allele specific assays tested. Results can be found in Table 5.2.

Figure 5.4 - Percentage of contigs containing SNPs between the resistant and susceptible genotypes in each Mercator mapping functional sub-category. FA: fatty acid; met.: metabolism; misc.: miscellaneous; PS: photosynthesis

167 Chapter 5 ______

5.4.7. EST-SRR markers development

EST-SSRs were identified through the Phobos software (Mayer, 2010), using as search parameters, perfect SSRs with a repeat unit length of two to six nucleotides. Polymorphisms between the partially resistant and susceptible genotypes were manually identified and flanked by primer pair using the Primer3 software (Untergasser et al., 2012). From the 341 EST-SSR developed and tested through PCR amplification on the two studied L. cicera genotypes, 251 produced an amplicon, being 206 (60.4%) polymorphic between accessions BGE008277 and BGE023542 and 45 (13.2%) monomorphic. 31 (9.1%) primers pairs produced a complex pattern and the remaining 59 (17.3%) primer pairs failed to produce any fragment. The developed EST-SSR marker primers and genotyping results are listed in Additional File 5.7.

5.5. Discussion

The present study provided the first genomic profile of L. cicera pathogen interaction. Although there are previous studies on this species focused on the response against stresses such as rust (Vaz Patto et al., 2009), powdery mildew (Vaz Patto et al., 2007), broomrape (Fernández-Aparicio et al., 2009; Fernández-Aparicio and Rubiales, 2010) and bacterial blight (Martín-Sanz et al., 2012), none of those were performed at molecular level. A valuable set of novel molecular tools and expression information on genes potentially involved in the partial resistance of L. cicera against rust is now available from the present study, which will be useful in future precision breeding approaches.

168 Table 5.2 - Allele-specific expression analysis results. Summary of the RT-qPCR results in comparison to the allele distribution in RNA-Seq results. ∞ - only ______expressed in BGE023542 or BGE008277.

BGE008277 inoculated BGE023542 inoculated SNP ΔCт Mean ΔCт Mean Reference SNP refer- SNP BLASTn variant RNA-Seq RNA-Seq SNP assay RNA-Seq RNA-Seq SNP assay assay (BGE023542 - (BGE023542 - assembly posi- BLAST hit ence assay e-value base counts counts reference counts counts reference variant BGE008277) BGE008277) contig tion base variant reference variant base (Ct reference variant base (Ct base reference base variant base base (Ct base base mean) base base mean) (Ct mean) mean) Glucan endo-1,3-beta-glucosidase a22544;181 127 3,00E-168 A G 10 0 36,96 - 0 162 - 31,91 [Medicago truncatula] ∞ ∞ polygalacturonase inhibiting protein a16587;204 389 0 C T 24 0 - - 3 186 29,91 34,97 [Pisum sativum] ∞ ∞ Cell division protease ftsH-like protein a12135;196 396 0 A T 23 0 33,92 31,77 1 149 33,92 - 2,15 - [Medicago truncatula]

a28870;97 856 unknown [Medicago truncatula] 0 G A 1 13 35,92 - 110 0 - - 0,40 -

a1871;383 342 Peroxidase [Medicago truncatula] 6,00E-150 T C 1 19 - - 264 0 - - - -

a1871;383 593 Peroxidase [Medicago truncatula] 6,00E-150 C G 35 2 30,42 29,50 0 318 30,42 29,50 -3,44 -1,55

Calmodulin-binding heat shock protein a42821;85 1102 6,00E-150 G A 8 0 30,01 35,19 3 38 30,72 30,96 0,71 -4,23 [Medicago truncatula]

glucan endo-1,3-beta-d-glucosidase a601;548 419 0 C G 25 164 32,17 24,17 1812 10 19,22 24,69 -12,95 0,52 [Cicer arietinum]

ATP-dependent RNA helicase [Medicago a23248;109 415 3,00E-97 A G 0 21 - 30,93 73 0 35,95 - - - truncatula]

NADP-dependent D-sorbitol-6-phosphate Chickling pea response to rust a2860;344 1236 0 G A 76 3 - - 0 163 - - - - dehydrogenase [Medicago truncatula]

probable ubiquitin-conjugating enzyme E2 a10029;259 1404 0 G A 33 0 23,94 25,71 0 101 23,35 24,99 -0,59 -0,71 24-like [Glycine max]

hypothetical protein MTR_5g063940 a19261;130 1328 0 A G 14 1 - - 0 48 35,92 - - - [Medicago truncatula]

a33467;117 1535 actin-related protein 2-like [Glycine max] 0 T C 25 0 28,17 35,92 0 70 36,91 29,49 8,74 -6,43

a4980;381 529 protein notum homolog [Glycine max] 3,00E-58 A G 2 194 33,66 24,99 187 2 25,96 31,88 -7,70 6,89

Hypersensitive reaction associated Ca2+- a1697;460 388 1,00E-86 C T 116 0 - - 0 137 - 26,89 - binding protein [Medicago truncatula] ∞ ethylene-overproduction protein 1-like a716;325 744 0 C T 317 3 28,46 35,89 1 292 35,89 29,28 7,42 -6,60 [Glycine max] 1 69 indole-3-acetic acid-amido synthetase a16062;87 264 0 C A 40 6 - 36,74 0 26 - 36,93 - 0,19 GH3.6-like [Glycine max]

Chapter 5 ______

The first set of specific EST-SSRs and SNP-based markers for this plant species was developed within this study. The performed EST-SSR markers validation confirmed 206 new EST-SSRs polymorphic between two L. cicera genotypes, which may support the future construction of a linkage map in this species, crucial to the development of more effective precision breeding approaches. With the SNP information, SNP base markers can be developed, such as the dual labelled probe RT-qPCR assays here presented, that can be used for eQTL studies. With the SNP data it is also possible the development of SNP arrays for diversity and mapping studies.

Incongruence between the expected and experimental results in the EST-SSR validation may be due to the fact that RNA-Seq data provides information only from the exons. Therefore, primer pairs that failed to amplify or amplified a complex pattern of fragments might be caused by the presence of large introns in the flanking region not detectable in the available RNA-Seq data. Also, not validated EST- SSRs may be due to the existence of homologous regions in others genetic regions, or the primers being located across splice regions. The lack of an amplicon could also be due to primers derived from chimeric cDNA clones (Varshney et al., 2005).

Allele-specific expression of candidate resistance genes points out to more complex levels in regulation of gene expression that can be further explored (Zhang and Borevitz, 2009). In our allele-specific expression RT-qPCR assays, a low number of assays corroborated the RNA-Seq data. One reason for the failure of many allele-specific assays is the selection of sites with partly quite low expression levels. In combination with limitations derived from assay design this is one explanation why several assays did not confirm the RNA-Seq data. An additional explanation for the failure of the majority of the assays is the

170 Chickling pea response to rust ______limited resolution of common qPCR systems like the StepOne applied in this study. These results were more evident since we mainly analysed alleles with low expression. We expect more positive results for the developed allele-specific assays in future applications, like eQTL mapping, using a Droplet Digital PCR system (ddPCR platform), that has a much higher resolution – e.g. five logs of dynamic range with the BIORAD Q100 system - what will increase especially the detection of rare alleles like ‘a16062;87’ with a maximum of 40 normalized reads only in RNA-Seq data (Table 5.2).

The lack of genomic resources is common in orphan crops, and L. cicera was not an exception. From all the contigs assembled in this study, only 42% were successfully BLASTed in plant databases. Since we sequenced samples inoculated with rust, we could also identify transcripts that probably have a fungal origin. Those were 1.2% of the total transcripts. As already discussed by Hacquard et al. (2011), the low number of fungal transcripts may reflect the low number of fungal structures in early-infected leaves.

This study was pioneer in providing a molecular overview of defence responses in L. cicera against U. pisi. U. pisi penetrates through the stoma, thus, pathogen perception is the first key element in defence response. Pathogen-associated molecular pattern triggered immunity (PTI) rely on an efficient signalling network in order to restrain infection (Nicaise et al., 2009). The receptor-like kinase 53 (RLP53), identified exclusively up-regulated in the partially resistant genotype, is a mitogen activated protein kinase (MAPK) associated with biotic stress signalling, that along with the WRKY transcription factor, also up-regulated exclusively in the partially resistant genotype, could be important in the defence response. WRKY transcription factors are well

171 Chapter 5 ______known elements that can influence pathogen perception (Eulgem, 2005).

Also related with signaling, several MLO-like transcripts were identified differentially expressed. The MLO gene was first identified in barley, where mutations in this gene were found to confer resistance to powdery mildew (Jørgensen, 1992) and up to now several MLO genes were identified as being responsible for susceptibility (Chen et al., 2006; Kim and Hwang, 2012; Zheng et al., 2013; McGrann et al., 2014). In L. cicera, a MLO-like transcript (AtMLO6, PsMLO1, a289255;11) was also detected up-regulated in the susceptible genotype and down regulated in the partially resistant genotype. This MLO-like gene was already identified as a susceptibility gene, mediating the vulnerability to several fungal pathogens in Arabidopsis (Chen et al., 2006) and to powdery mildew in Pisum sativum (Humphry et al., 2011). Recently, this gene was found up-regulated in a partially resistant L. sativus genotype, while in the resistant genotype the transcript for this gene was not detected (Almeida et al., 2014b). These findings ensure the importance of AtMLO6/PsMLO1 homologs as susceptibility genes in several Lathyrus spp. what should be further explored in the future.

After entering the mesophyll through the stoma, the rust fungal hyphae try to penetrate the mesophyll cells where haustoria are formed (Vaz Patto and Rubiales, 2009). The perception of cell wall modifications, through components released through cell wall degradation by the pathogen, can activate plant local responses triggering repair and fortification mechanisms by the expression of different genes (Cantu et al., 2008). Four transcripts described as being involved in cell wall degradation were identified exclusively up- regulated in the partially resistant genotype. The product of these

172 Chickling pea response to rust ______genes may act directly in the pathogen cell wall, having an antimicrobial activity, or act in the own plant cell wall producing damage-associated molecular patterns (DAMPs) to activate defence responses (Boller and Felix, 2009).

Also three cellulose synthase ‘IRREGULAR XYLEM 1 (IRX1)’, involved in cell wall synthesis (Taylor et al., 2000), were identified as being down-regulated in both genotypes upon inoculation with rust (groups F, G and H). IRX1 was previously found up-regulated in a L. sativus resistant genotype also in response to rust infection (Almeida et al., 2014b). This evidence indicate that the induction of this cellulose synthase, and consequent cell wall strengthening, may play an important role in the diverse resistance mechanisms presented by this L. sativus genotype and absent in L. cicera genotypes.

Another gene involved in cell wall synthesis, identified as down- regulated (one transcript down-regulated in both genotypes and another transcript down-regulated only in the susceptible genotype) was the ‘cellulose synthase 3 (CESA3)’. CESA3-deficient Arabidopsis mutants showed to have reduced levels of cellulose synthesis, which activated lignin synthesis and defence responses through the jasmonate and the ethylene signaling pathways (Caño-Delgado et al., 2003).

Related with secondary metabolism, two transcripts encoding the ‘3-ketoacyl-CoA synthase 6 (KCS6)’, involved in the biosynthesis of very long chain fatty acid (VLCFA) were identified exclusively up- regulated in the partially resistant genotype. The VLCFAs are fatty acids synthesized in the endoplasmic reticulum, which are crucial for a wide range of biological processes in plants, including defence. These lipids are known to be required for the biosynthesis of the plant cuticle

173 Chapter 5 ______

(Samuels et al., 2008), and the generation of sphingolipids (Worrall et al., 2003), that can play a direct role in defence (Berkey et al., 2012).

Also exclusively up-regulated in the partially resistant genotype (group E) a ‘DOWNY MILDEW RESISTANT 6 (DMR6) was identified. DMR6 was found to encode a 2-oxoglutarate (2OG)-Fe(II) oxygenase of unknown function. Despite its name, the expression of this gene is required for susceptibility to Hyaloperonospora parasitica and Colletotrichum higginsianum in Arabidopsis (Van Damme et al., 2008).

Peroxidases localised in the cell wall generate hydrogen peroxide, contributing as a source of reactive oxygen species (Passardi et al., 2005; Daudi et al., 2012), that leads to oxidative stress, a serious imbalance between production of ROS and antioxidant defences. In plants, the balance between ROS production and antioxidant defence determines the extent of oxidative damage (Moller et al., 2007). Since no hypersensitive response was detected in the L. cicera response to U. pisi infection, we may assume that the only peroxidase identified exclusively up-regulated in the partially resistant genotype is insufficient to produce a visible hypersensitive response.

The resistance to rust depicted by several L. cicera and L. sativus genotypes is due to a restriction of haustoria formation with high percentage of early abortion of colonies, with an associated reduction of number of haustoria per colony and reduction of intercellular growth of infection hyphae (Vaz Patto et al., 2009; Vaz Patto and Rubiales, 2009; Vaz Patto and Rubiales, 2014). Although the resistance mechanisms against rust infection are the same, those L. sativus and L. cicera genotypes present on average different resistance levels, being the most resistant L. cicera genotype (DS = 36%) (Vaz Patto et al., 2009) slightly more susceptible than the most susceptible L. sativus genotype (DS = 30%) (Vaz Patto and Rubiales, 2009).

174 Chickling pea response to rust ______

By comparing common transcripts between the presently obtained L. cicera results and the L. sativus transcriptomic profile response to the same rust (U. pisi) inoculation (Almeida et al., 2014b), 13 transcripts with contrasting expression patterns where identified. Two transcripts were exclusively up-regulated in both L. cicera genotypes and down regulated in both L. sativus genotypes, and 11 transcripts down-regulated in L. cicera and up-regulated in L. sativus. In this way an ‘acyl-CoA N-acetyltransferase (NAT) family protein’ (a330638;13) and a ‘Myristoyl-acyl carrier protein thiosterase (EC 3.1.2.-)’ (a39652;86) were found within the exclusively up-regulated genes in L. cicera.

Transcripts exclusively up-regulated in L. cicera (in general more susceptible) and exclusively down-regulated in L. sativus (in general more resistant) were involved in the jasmonic acid pathway, aromatic amino acid synthesis, flavonoid biosynthesis, stress abiotic, transport and signalling. Interestingly in signalling, the transcript identified corresponds to the ‘AtMLO8 (a228904;23)’, a gene described to be induced by wounding in Arabidopsis, but non-responsive to inoculation with the biotrophic fungi Erysiphe cichoracearum and Erysiphe orontii, the hemibiotrophic fungus Phytophtora infestans or the necrotrophic fungus Botrytis cinerea (Chen et al., 2006). It would be interesting to further study this gene in order to access its role in L. cicera higher susceptibility to rust.

When focusing on the most contrasting studied Lathyrus genotypes, L. sativus resistant genotype (DS=9%) and L. cicera susceptible genotype (DS=80%), nine transcripts were detected exclusively up-regulated in the resistant L. sativus genotype and exclusively down-regulated in the susceptible L. cicera genotype. Two of these transcripts related with signalling, three with transport and one

175 Chapter 5 ______with the phenylpropanoid/isoflavonoid pathways. In detail, in signalling, two receptor kinases were identified, a ‘mitogen-activated protein kinase kinase kinase 5 (MAPKKK5)’ (a271743;21) and a ‘receptor serine/threonine kinase with thaumatin domains’ (a130472;54). These genes might be involved on the pathogen perception and subsequent signalling cascades. Interestingly thaumatin is a pathogenesis related (PR) protein involved in increasing the permeability of fungal membranes by pore-forming mechanisms and therefore restraining fungal growth or even killing it (Selitrennikoff, 2001). In the functional category transport, a ‘cyclic nucleotide gated channel 1 (CNGC)’ (a61456;73) and two ‘aminophospholipid ATPase 1 (ALA1)’ (a162161;22 and a209230;31) were identified. CNGCs were found to be involved in the defence responses of Arabidopsis to inoculation with P. syringae and H. parasitica, as reviewed by Kaplan et al (2007), while ALA1 has been associated with cold tolerance, potentially involved in generating membrane lipid asymmetry (Gomes et al., 2000). Finally the ‘Isoflavone-7-O-methyltransferase 9 (IOMT9) (EC 2.1.1.150)’ (a27045;83), was identified up-regulated in the resistant L. sativus genotype while being down-regulated in the L. cicera susceptible genotype. It is reported that the overexpression of IOMTs in Medicago sativa, induce the phenylpropanoid/isoflavonoid pathway, conferring resistance to Phoma medicaginis (He and Dixon, 2000).

In order to access transcripts contributing to the overall resistance in Lathyrus spp. we compared transcripts from the most resistant genotypes, L. sativus resistant genotype (BGE015746; DS=9%), L. sativus partially resistant genotype (BGE024709; DS=30%) and L. cicera partially resistant genotype (BGE023542; DS=36%) against the susceptible L. cicera genotype (BGE008277; DS=80%). For that, common transcripts were filtered using the

176 Chickling pea response to rust ______following conditions: exclusively up-regulated in L. sativus genotypes, exclusively up-regulated in the partially resistant L. cicera genotype and exclusively down-regulated in the susceptible L. cicera genotype. A total of twenty five transcripts met these criteria. From those, six transcripts are known to be associated with plant defence responses, namely two transcripts encoding for ‘PENETRATION 3 (PEN3)’ (a14741;145 and a49947;119), a gene required for non-host penetration resistance to Blumeria graminis and Phytophtora infestans in Arabidopsis, being this gene also required for mlo-mediated resistance (reviewed by Hückelhoven, 2007). Another transcript identified was the ‘ethylene response factor 5 (ERF5)’ (a259652;14), an inducer of the SA biosynthesis pathway that also inhibits the JA and ET biosynthesis pathways (Son et al., 2011), an expected pattern in response to biotrophs (Bari and Jones, 2009). Also identified with this pattern was a transcript encoding for a ‘glucan endo-1,3-beta glucosidase (EC 3.2.1.39)’ (a44607;146). Beta-glycosidases are involved in diverse and important functions in plants, including bioactivation of defence compounds, cell wall degradation in endosperm during germination, activation of phytohormones, and lignifications (Morant et al., 2008). Moreover an ‘acidic endochitinase precursor (EC 3.2.1.14)’ (a21834;176) was also identified in the same pattern. Chitinases are involved in the inhibition of fungal hyphae growth in intercellular spaces as a defence response to fungal infection in several plant species (Grover, 2012). Finally, also identified with this pattern a ‘disease resistance family protein/leucine-rich repeat family protein’ (a26409;113), potentially involved in disease resistance which function remains unknown and representing an interesting candidate for further studies.

177 Chapter 5 ______

5.6. Conclusions

Our results provided an overview of gene expression profiles of contrasting L. cicera genotypes inoculated with rust, suggesting a different regulation of genes involved in signalling, cell wall metabolism and in the synthesis of secondary metabolites as the genetic basis of partial resistance to rust. The differentially expressed genes identified may be significant for the establishment or prevention of infection. Those are suitable candidate genes for future functional studies in order to shed light on the molecular mechanisms of plant-pathogen interactions.

The L. cicera gene expression results along with the previous information obtained from the L. sativus transcriptome inoculated with the same pathogen (Almeida et al., 2014b), offered a valuable set of sequence data for candidate rust resistant gene discovery in this genus. The design of specific EST-SSRs and detection and validation of SNPs for the first time on L. cicera will support future genetic studies on this species. These new molecular tools are suitable for studies involving marker-trait association, QTL and eQTL mapping and genetic diversity analysis.

5.7. Acknowledgments

NFA carried out the inoculations and sample processing for sequencing and RT-qPCR, designed the primers, performed and analysed the RT-qPCR experiments, selected the SNPs for the allele- specific expression, developed the EST-SSRs primers and performed the all the data analysis and drafted the manuscript. RH designed and tested the dual labelled probes. STL and TAF performed the EST-SSR genotyping. NK and BR performed the RNA sequencing and

178 Chickling pea response to rust ______bioinformatics data processing. PW and DR revised the manuscript critically. MCVP coordinated the study and participated to the drafting and revision of the manuscript. NFA would like to thank the CRF-INIA, Madrid, Spain, for supplying the genotypes.

5.8. References

Almeida, N.F., Leitão, S.T., Caminero, C., Torres, A.M., Rubiales, D., and Vaz Patto, M.C. (2014a). Transferability of molecular markers from major legumes to Lathyrus spp. for their application in mapping and diversity studies. Molecular Biology Reports 41, 269-283. Almeida, N.F., Leitão, S.T., Krezdorn, N., Rotter, B., Winter, P., Rubiales, D., and Vaz Patto, M.C. (2014b). Allelic diversity in the transcriptomes of contrasting rust-infected genotypes of Lathyrus sativus, a lasting resource for smart breeding. BMC Plant Biology 14, 376. Andersen, C.L., Jensen, J.L., and Orntoft, T.F. (2004). Normalization of real-time quantitative reverse transcription-PCR data: a model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Res 64, 5245-5250. Bari, R., and Jones, J.G. (2009). Role of plant hormones in plant defence responses. Plant Molecular Biology 69, 473-488. Barilli, E., Moral, A., Sillero, J.C., and Rubiales, D. (2012). Clarification on rust species potentially infecting pea (Pisum sativum L.) crop and host range of Uromyces pisi (Pers.) Wint. Crop Protection 37, 65-70. Barilli, E., Sillero, J.C., Serrano, A., and Rubiales, D. (2009). Differential response of pea (Pisum sativum) to rusts incited by Uromyces viciae-fabae and U. pisi. Crop Protection 28, 980- 986. Berkey, R., Bendigeri, D., and Xiao, S. (2012). Sphingolipids and plant defense/disease: The “death” connection and beyond. Frontiers in Plant Science 3, 68. Boller, T., and Felix, G. (2009). A Renaissance of elicitors: Perception of microbe-associated molecular patterns and danger signals by pattern-recognition receptors. Annual Review of Plant Biology 60, 379-406. Brem, R.B., Yvert, G., Clinton, R., and Kruglyak, L. (2002). Genetic dissection of transcriptional regulation in budding yeast. Science 296, 752-755.

179 Chapter 5 ______

Caño-Delgado, A., Penfield, S., Smith, C., Catley, M., and Bevan, M. (2003). Reduced cellulose synthesis invokes lignification and defense responses in Arabidopsis thaliana. The Plant Journal 34, 351-362. Cantu, D., Vicente, A.R., Labavitch, J.M., Bennett, A.B., and Powell, A.L. (2008). Strangers in the matrix: plant cell walls and pathogen susceptibility. Trends in Plant Science 13, 610-617. Chen, Z., Hartmann, H.A., Wu, M.-J., Friedman, E., Chen, J.-G., Pulley, M., Schulze-Lefert, P., Panstruga, R., and Jones, A. (2006). Expression analysis of the AtMLO gene family encoding plant- specific seven-transmembrane domain proteins. Plant Molecular Biology 60, 583-597. Consonni, C., Humphry, M., and Panstruga, R. (2009). "Defence Responses in Plants," in Plant Relationships, ed. H. Deising. (Berlin Heidelberg, Germany: Springer), 363-385. Cowles, C.R., Hirschhorn, J.N., Altshuler, D., and Lander, E.S. (2002). Detection of regulatory variation in mouse genes. Nature Genetics 32, 432-437. Daudi, A., Cheng, Z., O’brien, J.A., Mammarella, N., Khan, S., Ausubel, F.M., and Bolwell, G.P. (2012). The apoplastic oxidative burst peroxidase in Arabidopsis is a major component of pattern- triggered immunity. The Plant Cell 24, 275-287. Drummond, A.J., Ashton, B., Buxton, S., Cheung, M., Cooper, A., Duran, C., Field, M., Heled, J., Kearse, M., Markowitz, S., Moir, R., Stones-Havas, S., Sturrock, S., Thierer, T., and Wilson, A. (2011). "Geneious v5.4", (Available from http://www.geneious.com/). Edgar, R.C. (2010). Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460-2461. Eulgem, T. (2005). Regulation of the Arabidopsis defense transcriptome. Trends in Plant Science 10, 71-78. Fernández-Aparicio, M., Flores, F., and Rubiales, D. (2009). Field response of Lathyrus cicera germplasm to crenate broomrape (Orobanche crenata). Field Crops Research 113, 321-327. Fernández-Aparicio, M., and Rubiales, D. (2010). Characterisation of resistance to crenate broomrape (Orobanche crenata Forsk.) in Lathyrus cicera L. Euphytica 173, 77-84. Gomes, E., Jakobsen, M.K., Axelsen, K.B., Geisler, M., and Palmgren, M.G. (2000). Chilling tolerance in Arabidopsis involves ALA1, a member of a new family of putative aminophospholipid translocases. The Plant Cell 12, 2441-2454. Grover, A. (2012). Plant Chitinases: Genetic Diversity and Physiological Roles. Critical Reviews in Plant Sciences 31, 57- 73.

180 Chickling pea response to rust ______

Guo, M., Rupe, M.A., Zinselmeier, C., Habben, J., Bowen, B.A., and Smith, O.S. (2004). Allelic variation of gene expression in maize hybrids. Plant Cell 16, 1707-1716. Hacquard, S., Veneault-Fourrey, C., Delaruelle, C., Frey, P., Martin, F., and Duplessis, S. (2011). Validation of Melampsora larici- populina reference genes for in planta RT-quantitative PCR expression profiling during time-course infection of poplar leaves. Physiological and Molecular Plant Pathology 75, 106- 112. Hanbury, C.D., Siddique, K.H.M., Galwey, N.W., and Cocks, P.S. (1999). Genotype-environment interaction for seed yield and ODAP concentration of Lathyrus sativus L. and L. cicera L. in Mediterranean-type environments. Euphytica 110, 45-60. He, X.Z., and Dixon, R.A. (2000). Genetic manipulation of isoflavone 7-O-methyltransferase enhances biosynthesis of 4'-O- methylated isoflavonoid phytoalexins and disease resistance in alfalfa. The Plant Cell 12, 1689-1702. Hopf, M. (1986). "Archaeological evidence of the spread and use of some members of the Leguminosae family," in The Origin and domestication of cultivated plants: symposium, ed. C. Barigozzi. (Oxford, U.K.: Elsevier), 35-60. Huang, X., and Madan, A. (1999). CAP3: A DNA sequence assembly program. Genome Research 9, 868-877. Hückelhoven, R. (2007). Cell wall–associated mechanisms of disease resistance and susceptibility. Annual Review of Phytopathology 45, 101-127. Humphry, M., Reinstadler, A., Ivanov, S., Bisseling, T., and Panstruga, R. (2011). Durable broad-spectrum powdery mildew resistance in pea er1 plants is conferred by natural loss-of-function mutations in PsMLO1. Molecular Plant Pathology 12, 866-878. Jørgensen, I.H. (1992). Discovery, characterization and exploitation of Mlo powdery mildew resistance in barley. Euphytica 63, 141- 152. Kaplan, B., Sherman, T., and Fromm, H. (2007). Cyclic nucleotide- gated channels in plants. FEBS Letters 581, 2237-2246. Kenicer, G.J., Kajita, T., Pennington, R.T., and Murata, J. (2005). Systematics and biogeography of Lathyrus (Leguminosae) based on internal transcribed spacer and cpDNA sequence data. American Journal of Botany 92, 1199-1209. Kim, D.S., and Hwang, B.K. (2012). The pepper MLO gene, CaMLO2, is involved in the susceptibility cell-death response and bacterial and oomycete proliferation. The Plant Journal 72, 843- 855.

181 Chapter 5 ______

Lohse, M., Nagel, A., Herter, T., May, P., Schroda, M., Zrenner, R., Tohge, T., Fernie, A.R., Stitt, M., and Usadel, B. (2014). Mercator: a fast and simple web server for genome scale functional annotation of plant sequence data. Plant Cell and Environment 37, 1250-1258. Marioni, J.C., Mason, C.E., Mane, S.M., Stephens, M., and Gilad, Y. (2008). RNA-seq: an assessment of technical reproducibility and comparison with gene expression arrays. Genome Research 18, 1509-1517. Martín-Sanz, A., Palomo, J., Pérez De La Vega, M., and Caminero, C. (2012). Characterization of Pseudomonas syringae pv. syringae isolates associated with bacterial blight in Lathyrus spp. and sources of resistance. European Journal of Plant Pathology 134, 205-216. Mayer, C. (2010). "Phobos 3.3.11, [http://www.rub.de/spezzoo/cm/cm_phobos.htm]".). Mcgrann, G.R., Stavrinides, A., Russell, J., Corbitt, M.M., Booth, A., Chartrain, L., Thomas, W.T., and Brown, J.K. (2014). A trade off between mlo resistance to powdery mildew and increased susceptibility of barley to a newly important disease, Ramularia leaf spot. Journal of Experimental Botany 65, 1025-1037. Moller, I.M., Jensen, P.E., and Hansson, A. (2007). Oxidative modifications to cellular components in plants. Annual Review of Plant Biology 58, 459-481. Morant, A.V., Bjarnholt, N., Kragh, M.E., Kjærgaard, C.H., Jørgensen, K., Paquette, S.M., Piotrowski, M., Imberty, A., Olsen, C.E., Møller, B.L., and Bak, S. (2008). The β-glucosidases responsible for bioactivation of hydroxynitrile glucosides in Lotus japonicus. Plant Physiology 147, 1072-1091. Nicaise, V., Roux, M., and Zipfel, C. (2009). Recent advances in PAMP-triggered immunity against bacteria: pattern recognition receptors watch over and raise the alarm. Plant Physiology 150, 1638-1647. Passardi, F., Cosio, C., Penel, C., and Dunand, C. (2005). Peroxidases have more functions than a Swiss army knife. Plant Cell Reports 24, 255-265. Roth, A., Ding, J., Morin, R., Crisan, A., Ha, G., Giuliany, R., Bashashati, A., Hirst, M., Turashvili, G., Oloumi, A., Marra, M.A., Aparicio, S., and Shah, S.P. (2012). JointSNVMix: a probabilistic model for accurate detection of somatic mutations in normal/tumour paired next-generation sequencing data. Bioinformatics 28, 907-913.

182 Chickling pea response to rust ______

Rubiales, D., Sillero, J.C., and Emeran, A.A. (2013). Response of vetches (Vicia spp.) to specialized forms of Uromyces vicia- fabae and to Uromyces pisi. Crop Protection 46, 38-43. Samuels, L., Kunst, L., and Jetter, R. (2008). Sealing plant surfaces: cuticular wax formation by epidermal cells. Annual Review of Plant Biology 59, 683-707. Schaefer, H., Hechenleitner, P., Santos-Guerra, A., Sequeira, M.M.D., Pennington, R.T., Kenicer, G., and Carine, M.A. (2012). Systematics, biogeography, and character evolution of the legume tribe Fabeae with special focus on the middle-Atlantic island lineages. BMC Evolutionary Biology 12, 250. Selitrennikoff, C.P. (2001). Antifungal proteins. Applied and Environmental Microbiology 67, 2883-2894. Son, G.H., Wan, J., Kim, H.J., Nguyen, X.C., Chung, W.S., Hong, J.C., and Stacey, G. (2011). Ethylene-Responsive Element-Binding Factor 5, ERF5, is involved in chitin-induced innate immunity response. Molecular Plant-Microbe Interactions 25, 48-60. Taylor, N.G., Laurie, S., and Turner, S.R. (2000). Multiple cellulose synthase catalytic subunits are required for cellulose synthesis in Arabidopsis. The Plant Cell 12, 2529-2539. Thimm, O., Blasing, O., Gibon, Y., Nagel, A., Meyer, S., Kruger, P., Selbig, J., Muller, L.A., Rhee, S.Y., and Stitt, M. (2004). MAPMAN: a user-driven tool to display genomics data sets onto diagrams of metabolic pathways and other biological processes. The Plant Journal 37, 914-939. Torres, A.M., Weeden, N.F., and Martin, A. (1993). Linkage among isozyme, RFLP and RAPD markers in Vicia faba. Theoretical and Applied Genetics 85, 937-945. Untergasser, A., Cutcutache, I., Koressaar, T., Ye, J., Faircloth, B.C., Remm, M., and Rozen, S.G. (2012). Primer3-new capabilities and interfaces. Nucleic Acids Research 40, e115. Van Damme, M., Andel, A., Huibers, R.P., Panstruga, R., Weisbeek, P.J., and Van Den Ackerveken, G. (2005). Identification of Arabidopsis loci required for susceptibility to the downy mildew pathogen Hyaloperonospora parasitica. Molecular Plant- Microbe Interactions 18, 583-592. Van Damme, M., Huibers, R.P., Elberse, J., and Van Den Ackerveken, G. (2008). Arabidopsis DMR6 encodes a putative 2OG-Fe(II) oxygenase that is defense-associated but required for susceptibility to downy mildew. The Plant Journal 54, 785-793. Varshney, R.K., Graner, A., and Sorrells, M.E. (2005). Genic microsatellite markers in plants: features and applications. Trends in Biotechnology 23, 48-55.

183 Chapter 5 ______

Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2007). Resistance reaction to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus cicera from Iberian origin. Genetic Resources and Crop Evolution 54, 1517-1521. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2009). Pre and posthaustorial resistance to rusts in Lathyrus cicera L. Euphytica 165, 27-34. Vaz Patto, M.C., and Rubiales, D. (2009). Identification and characterization of partial resistance to rust in a germplasm collection of Lathyrus sativus L. Plant Breeding 128, 495-500. Vaz Patto, M.C., and Rubiales, D. (2014). Resistance to rust and powdery mildew in Lathyrus crops. Czech Journal of Genetics and Plant Breeding 50, 116-122. Vaz Patto, M.C., Skiba, B., Pang, E.C.K., Ochatt, S.J., Lambein, F., and Rubiales, D. (2006). Lathyrus improvement for resistance against biotic and abiotic stresses: From classical breeding to marker assisted selection. Euphytica 147, 133-147. Wang, L., Feng, Z., Wang, X., and Zhang, X. (2010). DEGseq: an R package for identifying differentially expressed genes from RNA-seq data. Bioinformatics 26, 136-138. Worrall, D., Ng, C.K., and Hetherington, A.M. (2003). Sphingolipids, new players in plant signaling. Trends in Plant Science 8, 317- 320. Yan, H., Yuan, W., Velculescu, V.E., Vogelstein, B., and Kinzler, K.W. (2002). Allelic variation in human gene expression. Science 297, 1143. Zhang, X., and Borevitz, J.O. (2009). Global analysis of allele specific expression in Arabidopsis thaliana. Genetics. Zheng, Z., Nonomura, T., Appiano, M., Pavan, S., Matsuda, Y., Toyoda, H., Wolters, A.-M.A., Visser, R.G.F., and Bai, Y. (2013). Loss of function in Mlo orthologs reduces susceptibility of pepper and tomato to powdery mildew disease caused by Leveillula taurica. PLoS ONE 8, e70723.

184 Chapter 6

Genetic basis of partial resistance to rust and powdery mildew in Lathyrus cicera - QTL mapping and synteny studies

The work presented in this chapter was mostly performed by Nuno Felipe Almeida (see Acknowledgements section), and corresponds to the following manuscript in preparation:

Almeida, N.F., Alves, M.L., Leitão, S.T., Rubiales, D., and Vaz Patto, M.C. Genetic basis of partial resistance to rust and powdery mildew in Lathyrus cicera - QTL mapping and synteny studies

185

Disease resistance QTL mapping and Synteny studies in L. cicera ______

6.1 Abstract

Lathyrus cicera L. (chickling pea) is a cool season fodder crop that has a great capacity to adapt to unfavourable environments (e.g., drought, flood and salinity). Besides its tolerance to abiotic factors, this species presents resistance to some important legume diseases like rust, powdery mildew, ascochyta blight or bacterial blight. However, no molecular tools exist to aid in the elucidation of the molecular mechanisms involved in the resistance process. For that, this work reports the first linkage map for L. cicera, and QTL mapping for rust and powdery mildew resistance, based in a RIL population. Additionally, the data enable comparative mapping between L. cicera and Medicago truncatula. The map was constructed with 258 molecular markers (21 EST-SSRs, 4 ITAPs and 233 SNPs), covering 757.11 cM of genetic distance organized on 8 major and 3 minor linkages groups, with an average distance between markers of 2.93 cM. The synteny study indicates a high macrosyntenic conservation between L. cicera and M. truncatula. One QTL for resistance to powdery mildew (explaining 31.4% of the phenotypic variance) and two QTLs for partial resistance to rust (explaining 26.8% of the phenotypic variance) were detected. QTL analysis revealed that the genetic control of the partial resistances to rust and powdery mildew was polygenic. These new genetic and genomic information represents the foundation stones for a more effective L. cicera breeding.

6.2. Introduction

Lathyrus cicera L. (chickling pea) is an annual temperate legume mainly grown as stock feed, both as forage and grain

187 Chapter 6 ______

(Hanbury et al., 1999). It can adapt well to harsh environments, being tolerant to drought, water lodging and resistant to several important legume pathogens. Due to that, L. cicera is considered a good alternative for low-input cropping systems in more marginal lands generally more prone to these biological stresses (Vaz Patto et al., 2006b). Rust (Vaz Patto et al., 2009), powdery mildew (Vaz Patto et al., 2007), bacterial blight (Martín-Sanz et al., 2012) and crenate broomrape (Fernández-Aparicio et al., 2009) are among the biotic constrains to which L. cicera presents different resistance levels.

Rusts are among the most important diseases of legumes (Sillero et al., 2006) affecting also Lathyrus spp. (Duke, 1981; Campbell, 1997; Vaz Patto et al., 2006b). Rusts are caused by biotrophic fungi that keep infected host cells alive, depending on the hosts to reproduce and complete their life cycles. Although some rusts can be cultured on very complex synthetic media, they have no known saprotrophic existence in nature (Staples, 2000). Rusts form elaborate intracellularlly accommodated structures called haustoria, that allow contact between fungal and plant cells over a prolonged period of time (O’Connell and Panstruga, 2006). In Lathyrus spp., rust is caused by Uromyces pisi (Pers.) Wint. and U. viciae-fabae (Pers.) J. Schröt. Both hypersensitive and partial resistance response to rust has been described in L. cicera. Partial resistance is due to a restriction of haustoria formation with high percentage of early aborted colonies, reduction of number of haustoria per colony and reduction of intercellular growth of infection hyphae (Vaz Patto et al., 2009; Vaz Patto and Rubiales, 2009; Vaz Patto and Rubiales, 2014).

Powdery mildews are probably the most common, conspicuous and widespread plant diseases, seldom killing their hosts, but utilizing their nutrients, reducing their photosynthesis and

188 Disease resistance QTL mapping and Synteny studies in L. cicera ______impairing growth, resulting in yield reductions up to 40% (Agrios, 2005). Powdery mildew in L. cicera is caused by Erysiphe pisi DC., that is the causal agent of pea powdery mildew. Erysiphe pisi is an obligate biotrophic ascomycete fungus characterized by its grey to white colonies formed on leaves, stem and pods of infected plants (Vaz Patto et al., 2006a) and can also be found on Medicago, Vicia, Lupinus, Lens and other Lathyrus spp. (Sillero et al., 2006). Pea powdery mildew is a serious disease of worldwide distribution, being particularly important in climates with warm, dry days and cool nights (Fondevilla and Rubiales, 2012). In general, the infection of L. cicera by pea powdery mildew presents a compatible reaction with no macroscopically visible necrosis. However, genotypes with reduced disease severity, despite of a high infection type, have been identified (Vaz Patto et al., 2006a; 2007; Vaz Patto and Rubiales, 2014), fitting the definition of partial resistance according to Parlevliet (1979). Until now, little is known about the genetic basis of rust and powdery mildew partial resistance in L. cicera. In the Chapter 5 of this thesis, the transcriptomic response of L. cicera to inoculation with rust was studied for the first time, indicating that the most differentially expressed transcripts, upon rust inoculation, between partially resistant (BGE023542) and susceptible (BGE008277) genotypes were involved in signalling, cell wall metabolism and synthesis of secondary metabolites.

Rusts and powdery mildews are airborne pathogens, with both sexual and asexual cycles of reproduction. This represents an increased complexity to the resistance trait stability, since resistance due to a single gene is more easily overcome by new races of the pathogens. In those cases, polygenic resistance, as in some cases of partial resistance, would be more difficult to overcome, and

189 Chapter 6 ______consequently is expected to be more durable (McDonald and Linde, 2002; Niks and Rubiales, 2002).

A linkage map of L. cicera would be crucial, although not yet available, to identify and locate the genes and genomic regions responsible to the resistance traits, opening the way for marked assisted selection (MAS) on this orphan species.

Molecular markers are the stepping stones in this mapping process. For L. cicera, a large set of expressed sequence tag - simple sequence repeat markers (EST-SSRs) and single nucleotide polymorphisms (SNPs) were already developed and may be used for mapping and diversity studies (Chapter 5). Also, other marker classes (intron-targeted amplified polymorphic (ITAP), EST-SSR, genomic simple sequence repeat (gSSR), resistance genes analogs (RGA) and disease resistance (DR) markers developed for other legume species (Medicago truncatula Gaertn., Pisum sativum L., Lens culinaris Medik., Lupinus spp. and Vicia faba L.) have been successfully cross-amplified in L. cicera (Almeida et al., 2014a). The existence of cross-species amplified markers would allow comparative mapping between L. cicera and related legume species, facilitating the exchange of genetic information in both directions.

In this study, and as a first step to understand the genetic basis of partial resistance to rust and powdery mildew in L. cicera, we used previously developed L. cicera molecular markers (described in Chapter 2 (Almeida et al., 2014a) and Chapter 5), to construct the first L. cicera linkage map. This approach has also allowed a comparative/synteny study of L. cicera with other legume species. This map was developed using a recombinant inbred line (RIL) population segregating for rust and powdery mildew resistance. This molecular information was then jointly analysed with the disease

190 Disease resistance QTL mapping and Synteny studies in L. cicera ______resistance scores of this population to study the genetic control of the L. cicera partial resistance to rust and powdery mildew.

6.3. Material and Methods

6.3.1. Plant Material

The mapping population used for the development of the linkage map consisted of 102 F5 Recombinant Inbred Lines (RILs) derived by single seed descendent from a cross between L. cicera genotypes BGE023542 and BGE008277. These two genotypes showed contrasting phenotypes to rust and powdery mildew infection and were kindly provided by the Plant Genetic Resources Centre (CRF-INIA), Madrid, Spain. BGE023542 shows partial resistance to rust, while BGE008277 is susceptible to rust and to powdery mildew (Vaz Patto et al., 2007; 2009).

6.3.2. Disease reaction evaluation

Disease reaction was evaluated under controlled conditions in three different periods. Rust evaluations were performed in the autumn of 2009, and in the spring of 2011 and 2013. Powdery mildew evaluations were performed in the autumn of 2009, spring of 2011 and autumn of 2011. These different evaluations will be referred in the manuscript as experiment 1R, 2R and 3R for rust and 1PM, 2PM and 3PM for powdery mildew.

6.3.2.1. Rust evaluation

The Uromyces pisi monosporic isolate UpCo-01, from the fungal collection of the Institute for Sustainable Agriculture-CSIC

191 Chapter 6 ______

(Córdoba, Spain), was used for the rust inoculation experiments. The inoculum was multiplied on plants of the very susceptible P. sativum cv. ‘Messire’ before use. Fifteen-day-old RILs seedlings, grown in plastic pots containing 250 cm3 of 1:1 sand-peat mixture in a controlled growth chamber (20 ± 2 ºC with a 12 h light photoperiod) were inoculated by dusting all the plants at the same time with 2 mg of spores per plant, diluted in pure talk (1:10), with the help of a small manual dusting device in a complete random experiment. Three replicates were used, each having four plants of each RIL individual family, the parental lines and two of cv. ‘Messire’ as control. Inoculated and control plants were incubated for 24 h at 20 ºC, in complete darkness, and 100% relative humidity, then transferred to the growth chamber and kept at 20 ± 2 ºC under 14 h light (150 μmol m−2 s−1) and 10 h dark. Rust reactions were assessed by measuring infection type (IT) and disease severity (DS). IT was scored 10 days after inoculation (d.a.i), and revised 15 d.a.i, using the IT scale of Stakman et al. (1962), where 0 = no symptoms, i = necrotic flecks, 1 = necrotic flecks with minute pustules barely sporulating, 2 = necrotic halo surrounding small pustules, 3 = chlorotic halo and 4 = well- formed pustules with no associated chlorosis or necrosis. Values 0–2 are considered indicative of resistance and 3–4 of susceptibility. DS was scored as the percentage of leaf area covered by rust pustules 15 d.a.i, on the second upper pair of open leaves, thus, the lower the DS value, the higher the resistance.

6.3.2.2. Powdery mildew evaluation

The Erysiphe pisi isolate CO-01, from the fungal collection of the Institute for Sustainable Agriculture-CSIC (Córdoba, Spain), was used for the powdery mildew inoculation experiments. The isolate

192 Disease resistance QTL mapping and Synteny studies in L. cicera ______derives from a population collected from an infected field (Córdoba, Spain) and maintained and multiplied before experiments on susceptible P. sativum cv. ‘Messire’ plants. RILs inoculation was done on detached L. cicera leaves from two-week-old seedlings using two leaves per each of six plants of each RIL individual family and were inoculated using a settling tower to give an inoculum density of about 20 conidia/mm2. Detached leaves were placed with the adaxial surface up on Petri dishes containing agar (4 g/L) + water + benzimidazole (62.5 mg/L). After inoculation, Petri dishes were covered and placed in the growth chamber at 20 ºC. Incubation started with a 6 h light period (250 µmol/m2) followed by a photoperiod of 14 h light and 10 h dark. Five d.a.i, the infection type and disease severity were measured. IT was recorded according to a 0–4 scale (Fondevilla et al., 2006), where 0 = no visible sign of disease, and 4 = well developed, freely sporulating colonies. DS was scored as the percentage of leaf coverage by the mycelium.

6.3.3. DNA isolation and molecular markers screening

Using fresh young leaves of one individual of each RIL family plus the two parental genotypes, DNA was extracted using a modified CTAB protocol developed by Torres et al. (1993). DNA was subsequently screened using different molecular markers. The markers used in this study were the following: Five heterologous ITAP markers selected from our previous work (Almeida et al., 2014a), identified as polymorphic and with an 1:1 Mendelian segregation in this mapping population; 57 EST-SSR makers predicted in silico from L. cicera parental genotypes RNA-Seq libraries (Chapter 5); 768 SNPs, selected taking into consideration their homology with the Medicago truncatula genome (MT3.5) (BLASTn; E-value < 1E-6) and

193 Chapter 6 ______their physical position in this genome to cover evenly M. truncatula’s chromosomic regions (http://www.medicagohapmap.org/tools/ blastsearch), preventing unwanted clustering of markers. Identification of the SSRs motifs and SNPs was performed as described in Chapter 5. PCR reactions and genotyping were performed as described in Almeida et al. (2014b), with the exception of SNP markers that were genotyped using an Illumina’s custom Golden Gate genotyping assay by Traitgenetics GmbH, Germany.

6.3.4. Map construction

Linkage analysis and segregation distortion tests were performed using JoinMap 4.0 software (van Ooijen, 2006), using a binary matrix including all the genotyping data as input. Markers with a severe segregation distortion (p ≤ 0.0005) were removed from the original molecular data set.

The determination of groups of linked markers (linkage groups) was done with a LOD score of 3. Linkage map calculations were done using all pairwise recombination estimates lower than 0.40 and a LOD score higher than 1.00, applying the Kosambi mapping function (Kosambi, 1943). The reliability of the obtained map was checked by inspecting the individual linkage group χ2 value.

6.3.5. Comparison with M. truncatula genome

Using the order of the SNP markers in the L. cicera linkage map and the information of the physical position of the same markers mapped on the M. truncatula genome (MT3.5) (BLASTn; E-value < 1E-6), were aligned in a matrix. Lines from the matrix correspond to the M. truncatula genome and the columns correspond to the L.

194 Disease resistance QTL mapping and Synteny studies in L. cicera ______cicera linkage groups that were rearranged in order to facilitate the visual estimation of co-linearity.

6.3.6. Phenotypic data analysis

In order to exclude outliers, RILs showing a DS value with standard deviation percentage higher than 30% were excluded from further analysis. Phenotypic data (disease reaction measurements) descriptive statistics analysis was done using SAS software (The SAS System for Windows version 9.2, Cary, NC, USA). Normality of residual distribution was checked using the Kolmogorov-Smirnov test. Spearman's correlation coefficients were computed for each trait between experiments by PROC CORR procedure. PROC GLM procedure was used for analysis of variance. In this model, environments (experiments) and genotypes were treated as fixed effects. Repetitions, treated as random, were nested in the environments. Genotype x Environment interaction was included in the model. The variance components for each trait in each environment where estimated using the PROC VARCOMP procedure. Broad-sense heritability, representing the part of the genetic variance in the total phenotypic variance, were calculated for each environment as: = /[ + ( / )] , where is the 2 2 2 2 2 genotypic variance, is ℎthe error𝛿𝛿𝑔𝑔 𝛿𝛿 𝑔𝑔variance𝛿𝛿 𝑟𝑟 and is the number𝛿𝛿𝑔𝑔 of 2 replications. 𝛿𝛿 𝑟𝑟

6.3.7. QTL mapping

The obtained linkage map was used for resistance to rust and powdery mildew QTL identification on this RIL population. Kruskal- Wallis single-marker analysis (non-parametric test), as well as both

195 Chapter 6 ______interval mapping (Lander and Botstein, 1989) and multiple-QTL mapping (MQM) (Jansen and Stam, 1994) were performed using MapQTL version 4.0 (Van Ooijen et al., 2002). A backward elimination procedure was applied to select cofactors significantly associated with each trait at P < 0.02 to be used in MQM. Genome- wide threshold values (P < 0.05) for declaring the presence of QTL were estimated from 10,000 permutations of each phenotypic trait (Churchill and Doerge, 1994). The 1-LOD and 2-LOD support intervals were determined for each LOD peak.

The R2 value, representing the percentage of the phenotypic variance explained by the marker genotype at the QTL, was taken from the peak QTL position as estimated by MapQTL. Additive effect for each detected QTL was estimated using the MQM procedure. Gene action was determined as described by Stuber et al. (1987) where: additive (|d/a| < 0.20); partial dominance (0.2 < |d/a| < 0.8); dominance (0.8 < |d/a| < 1.2); and overdominance (|d/a| > 1.20), where, d/a = dominance effects/additive effects. Linkage groups were drawn using MapChart version 2.2 software (Voorrips, 2002). Molecular markers were previously annotated via BLASTX as described in Chapter 5, allowing the identification of candidate resistance genes among the molecular markers within the QTL confidence intervals.

6.4. Results

6.4.1. Disease evaluation under controlled conditions The L. cicera parental genotypes and RILs used in this study presented an IT=4 for both diseases evaluated. BGE023542 displayed partial resistance to both rust and powdery mildew, with

196 Disease resistance QTL mapping and Synteny studies in L. cicera ______fully compatible interaction to both pathogens in spite of some reduced disease progress, with an average DS of 36% against the 52% observed in BGE008277 for rust evaluation across experiments, and DS of 33% for powdery mildew against the 73% observed in BGE008277. L. cicera RIL population rust and powdery mildew DS levels did not follow a normal distribution (Figure 6.1). An arcsine transformation was applied to DS to improve homogeneity of residual variance but no improvement in the normality of the data was observed. Positive transgressive segregation was detected for both disease reactions (Figure 6.1), with a higher degree for rust resistance, with more than half of the RIL families showing lower DS than the more resistant parent. No correlation between individual experiments for rust DS, and a weak correlation for powdery mildew DS, ranging from 0.19 (1PM vs. 2PM) and 0.25 (1PM vs. 3PM), were detected (Spearman’s correlation tests). Also, no correlation was detected when comparing the DS of the RILs to rust and powdery mildew inoculation (r=0.22).

Figure 6.1 - Frequency distributions of rust (experiment 2R) and powdery mildew (experiment 3PM) DS in the RIL families under controlled conditions. The average values of rust and powdery mildew DS of the two parental lines are indicated by the arrows.

197 Chapter 6 ______

The calculated broad-sense heritability for rust DS had a maximum value of 86% for experiment 3R, while for powdery mildew DS the maximum value was for experiment 2PM with 77%. Results for all the experiments can be found in Table 6.1.

Significant differences for rust and powdery mildew disease severity were detected between genotypes and the different experiments. The interaction between Genotype x Experiment was also significant, therefore, QTL mapping was performed separately for each experiment.

Table 6.1 – Phenotypic values (mean ± standard deviation) of the RIL families and quantitative genetic parameters for rust and powdery mildew resistance. Exp: experiment; Significance of the sources of variability: G-Genotype, E-Environment, Rep (E)-Repetitions within Environment, G x E-Genotype x Environment; Interaction-Levels of significance:ns non- significant value; *** significant at P < 0.001 RILs disease severity (%) Pearson's correlation Heritability (h2) (%) ANOVA (mean ± sd) ( r ) Exp. Exp. Exp. Exp. Exp. Exp. Rep Trait Exp. 1 Exp. 2 Exp. 3 G E G x E 1&2 1&3 2&3 1 2 3 (E) 30,0 ± 40,3 ± 20,0 ± rust 0,03 -0,02 -0,02 65.16 77.35 72.78 *** *** ns *** 10,1 11,3 4,4 powdery 60,9 ± 63,9 ± 46,6 ± 0,19 0,25 0,24 51.28 46.13 86.31 *** *** ns *** mildew 13,0 14,9 12,7

6.4.2. Linkage map construction

A total of 830 molecular markers (57 EST-SSRs, 5 ITAPs and

768 SNPs) were screened in the parental genotypes of the F5 RIL population. From these, 258 polymorphic loci were successfully mapped, namely, 21 EST-SSRs, 4 ITAPs and 233 SNPs.

In more detail, from the initial 57 EST-SSRs tested in the parental genotypes, 12 were excluded from the RILs genotyping because: four did not amplify any band, one presented a complex pattern and seven were monomorphic between the parental genotypes. From the 44 EST-SSRs screened in the RILs, 29 were

198 Disease resistance QTL mapping and Synteny studies in L. cicera ______selected for mapping, 28 with the expected Mendelian segregation (1 degree of freedom; α = 0.05; χ2 < 3.84) and one with a low segregation distortion (χ2 = 4.46). The remaining markers presented a high segregation distortion and were excluded from further analysis. One of the ITAPs presented a high segregation distortion and was removed. From the 768 SNPs genotyped in the RIL population, 369 were withdrawn from the mapping due to several causes: 206 for having more than 20% missing values and 163 with heterozygous individuals (not expected on a RIL population). A total of 399 SNPs were then selected for developing the linkage map. However, from the 399 SNP markers selected, 23 presented a severe segregation distortion (p ≤ 0.0005) and were also removed. From the remaining 376 markers, 81 were removed for being identical to other markers, thus redundant for the map construction. Additionally one RIL was removed because it had more than 25% of missing data.

Finally, the L. cicera linkage map was developed using data from 101 RILs screened with 258 polymorphic loci, 21 EST-SSRs, 4 ITAPs and 233 SNP. It covered 757.11 cM of genetic distance organized on 8 major and 3 minor linkages groups, with an average distance between markers of 2.93 cM. Only one marker could not be linked with any other LG. Twenty two percent of the markers showed significant deviation from the expected 1:1 segregation ratio (segregation distortion). A chromosomal region was considered skewed when four or more closely linked markers showed significant segregation distortion in the same direction (Xu et al., 1997). In the present linkage map these were observed in the extremity of linkage groups (LG) IV, V and IX, and in the centre of LG I and III. The smallest LGs (X and XI) were constituted exclusively by markers with high segregation distortion (Figure 6.2).

199 Chapter 6 ______

Inspection of the individual linkage group χ2 values for goodness-of-fit gave insight into the reliability of the obtained map. The χ2 values for all the linkage groups were < 1 (Table 6.2).

Table 6.2 - Description of the obtained linkage groups. Linkage number χ2 LOD Length Average Largest gap group of loci mean threshold (cM) distance (cM) (cM) I 46 0.174 4 150.16 3.26 15.05 II 41 0.133 4 134.40 3.28 14.80 III 47 0.417 4 114.61 2.44 11.73 IV 30 0.669 4 88.55 2.95 10.42 V 22 0.216 4 72.92 3.32 14.36 VI 24 0.234 4 66.64 2.78 18.61 VII 14 0.070 4 46.89 3.35 16.10 VIII 19 0.373 4 44.60 2.35 9.23 IX 9 0.221 3 25.36 2.82 6.27 X 3 0.025 4 9.22 3.07 8.65 XI 3 0.002 3 3.77 1.26 2.71

6.4.3. Macrosynteny between L. cicera LGs and M. truncatula chromosomes

Clear evidence of a simple and direct macrosyntenic relationship between the L. cicera and M. truncatula genome was detected in the dot matrix in Figure 6.3. The clear isoclinic diagonal line along the linkage groups provides a strong indication of the conservation of gene order in the two legume genomes. However, chromosomal rearrangements were also evident at a moderate level. For example, M. truncatula chromosomes 2 and 6 merged to form the L. cicera LG IV. Similarly, M. truncatula chromosome 4 splits into L. cicera LGs II and IX and M. truncatula chromosome 7 into L. cicera LGs VI and VII (Figure 6.3). Additionally, M. truncatula chromosome 8

200 Disease resistance QTL mapping and Synteny studies in L. cicera ______spans L. cicera LGs VIII, X and a large portion on the extremity of LG II.

Figure 6.2 - L. cicera first genetic linkage map based on a RIL population. Genetic distances given in cM (Kosambi mapping function) on the left. Marker names are shown on the right side of each linkage group with connecting lines indicating the position of each marker on the linkage group. Markers with distorted segregation ratios are marked with asterisks according to their significance levels (*: 0.05, **: 0.01, ***: 0.005, ****: 0.001, *****: 0.0005).

201 Chapter 6 ______

Figure 6.3 - Matrix plot of common gene-based SNP markers mapped in L. cicera and M. truncatula. The L. cicera and M. truncatula loci are listed horizontally and vertically, respectively, according to their linkage group order.

6.4.4. QTL mapping for rust and powdery mildew resistance

Two QTLs for partial resistance to rust (qrustres1 and qrustres2) and one to powdery mildew (qpmres1) were identified. In linkage group (LG) II, qrustres1 on experiment 2R, and qpmres1, on experiment 3PM (Figure 6.4) were detected. These QTLs explained, respectively, 14.1 and 31.4% of the phenotypic variance observed. Also in the rust experiment 2R, qrustres2 was detected in LG II explaining 12.7% of the phenotypic variance observed (Table 6.3). Resistant alleles (the ones presenting low DS values) were derived from the more resistant genotype for all QTLs. Using the estimated

202 Disease resistance QTL mapping and Synteny studies in L. cicera ______

additive effects of the BGE023542 alleles, we predicted a difference of 28.7% for rust resistance level between the two parental genotypes, based on simple additive model (two-times the sum of the additive affects). The observed phenotypic difference between the two parental genotypes was 25% for rust experiment R2.

Table 6.3 – Quantitative trait loci for rust and powdery mildew resistance in a L. cicera RIL5 population (BGE023542xBGE008277). Peak Additive Linkage Peak Marker Flanking R2 Trait QTL LOD effect Group (Positiona/LOD) markersb (%)d score (a)c Reaction c4_a35535 c4_a65394/ to rust qrustres1 II 3.27 -3.37 14.1 (55.793/3.27) c4_a26111 infection c4_a5026 c4_a17517/ qrustres2 II 2.97 -3.80 12.7 (115.603/2.97) c4_a1108 Reaction to c4_a12255 c5_a35535/ powdery qpmres1 II 6,71 -6.72 31.4 (58.658/6.71) c5_a5292 mildew infection a QTL position in cM from the top of the chromosome b molecular markers flanking the support interval estimated at a LOD fall of -2.00 c Additive effect = (mu_BGE023542 – mu_BGE008277) / 2; negative values indicate that the BGE023542 allele increased resistance trait value mu_BGE023542 – the estimated mean of the distribution of the quantitative trait associated with the BGE023542 allele mu_ BGE008277 – idem for the BGE008277 allele d Percent explained phenotypic variance.

6.4.5. Discovery of potential candidate genes for the identified QTLs

All the molecular markers present in this study were previously BLASTed against public plant protein databases (Chapter 5) in order to facilitate the identification of potential candidate genes, for rust and powdery mildew resistance QTLs. Two potential candidate genes were identified for the powdery mildew resistance QTL located on LG II (“Dihydroorotase” and “Protein trichome birefringence-like 27”). For the rust resistance QTLs

203 Chapter 6 ______located on LG II, one potential candidate gene was identified per QTL (qrustres1 and qrustres2), although only one of those two (“MAG2- interacting protein 3”) on qrustres1 was functionally annotated (Table 6.4). Exploring the homologous region from the identified QTLs in the M. truncatula genome, we observed that one of the PsMLO1 homologs was identified in the same region as the qpmres1 syntenic region.

Figure 6.4 - MQM QTL mapping analysis for rust and powdery mildew resistance on LG II of the L. cicera linkage map developed. Dashed line representing the significant LOD threshold for detecting QTLs.

204 ______

Table 6.4 – Potential candidate genes located under the identified in QTLs. Disease resistance QTL mapping and Synteny studies in L. cicera Linkage Markers QTL BLAST hit e-value Function Reference group under QTL

Dihydroorotase - involved in pirimidine metabolism (Giermann et al., 2002) qpmres1 II a12255;147 0 [Arabidopsis thaliana] - involved in production of secondary metabolism (Brown and Turan, 1995)

Protein trichome qpmres1 II a26111;129 birefringence-l ke 27 2.00E-22 - cellulose biosynthesis in Arabidopsis (Bischoff et al., 2010) [Arabidopsis thaliana]

- transport of proteins between ER and Golgi in MAG2-interacting protein 3 (Li et al., 2013) qrustres1 II a35535;94 0 Arabidopsis [Arabidopsis thaliana] (Yamada et al., 2011) - role of ER in defence though protein trafficking

hypothetical protein qrustres2 II a5026;279 MTR_4g019060 7.88E-179 - - [Medicago truncatula] qrustres1 – Rust resistance QTL 1 qrustres2 – Rust resistance QTL 2 qpmres1 – Powdery mildew resistance QTL 20 5

Chapter 6 ______

6.5. Discussion

In order to understand the genetic basis of the rust and powdery mildew partial resistance in L. cicera we performed a QTL analysis of these traits using a segregating RIL population. Resistance to both diseases was based on the reduction of disease severity, with parental genotypes and evaluated RILs showing a compatible interaction, with no associated cell death, fitting the definition of Partial Resistance sensu Parlevliet (1979).

To localize these resistance QTLs, we developed the first linkage map for L. cicera based on a RIL population. This linkage map is based on different types of molecular markers, including EST- SSRs, SNPs and ITAPs. A total of 1,113 molecular markers were screened in 102 individuals from a F5 RIL population segregating for rust and powdery mildew response to infection. The obtained map covered a total of 757.11 cM, with an average density of one marker every 2.93 cM, organized in 11 linkage groups, eight longer than 40 cM and 3 shorter groups.

Regions showing distorted segregation of molecular markers were found in clusters, mainly in the extremities of linkage groups. Also the two smaller LGs (X and XI) were exclusively constituted by segregation distorted markers, what might have contributed for their unlinked situation. Skewed regions were distorted towards the same direction (an excess of female parental line alleles), with the exception of the region on LG III, distorted towards an excess of male parental line alleles. These distorted regions may include potential lethal genes (Cheng et al., 1998), that when in homozygosity produce a lethal phenotype, and therefore, RILs containing those alleles will be absent from the mapping population. This will contribute for the

206 Disease resistance QTL mapping and Synteny studies in L. cicera ______segregation distortion of the markers linked to those genes that will not segregate independently from each other (Torjek et al., 2006).

A strong extensive co-linearity was detected between the L. cicera linkage groups and the M. truncatula genome. Similar high levels of marker order conservation have also been reported between M. truncatula and P. sativum (Aubert et al., 2006) and other closely related legumes such as M. truncatula and M. sativa (Choi et al., 2004a), L. culinaris and M. truncatula (Phan et al., 2007) and L. culinaris and P. sativum (Weeden et al., 1992). In the present study, a few rearrangements in the homologous marker order have also been detected, such as inversions (LG II and III) and translocations (LG II and IV). In a previous study comparing M. sativa and P. sativum synteny (Kaló et al., 2004), one rearrangement event was similar to the one observed in this study: in L. cicera, a linkage group (LG IV) was split in two, spanning chromosomes 2 and 6 in M. truncatula. In Kalo et al. (2004), P. sativum’s LG IV contained M. sativa’s LG 2 and LG 6, being acceptable to hypothesize that the chromosomal reduction between M. truncatula/M. sativa (n = 8) and L. cicera/P. sativum (n = 7) was caused by the fusion of M. truncatula chromosomes 2 and 6 (Choi et al., 2004b).

QTL analysis revealed that the genetic control of partial resistance to rust and powdery mildew was indeed polygenic taking into consideration the small percentage of total phenotypic variation explained by the detected QTLs. The QTL detected for resistance to powdery mildew explained 31.4% of the phenotypic variation and the two detected QTLs for partial resistance to rust explained a total of 26.8% of the phenotypic variation. The main advantage of the polygenic resistance is that it is generally more durable than monogenic resistance, that due to the higher selective pressure

207 Chapter 6 ______inflicted by the host to the pathogen, are more easily overcome by the fast evolving fungi (McDonald and Linde, 2002; Niks and Rubiales, 2002). The partial resistance reaction found in L. cicera is similar to what is found commonly in cool season legumes interaction with rusts and powdery mildews (Sillero et al., 2006; Rubiales et al., 2011), but due to the lack of anchor markers, our study was still unable to bridge the information with the other existing legume QTL resistance studies.

The QTL for powdery mildew resistance accounted for 31.4% of the total estimated phenotypic variance at seedling stage, while the two detected QTLs for rust resistance explained 14.1% and 12.7% of phenotypic variance. The remaining and unexplained variance may be due to other unidentified loci, which have not been detected either because of insufficient genome coverage or/and especially because of uncontrolled experimental and environmental variation. For instance, despite the use of the same U. pisi and E. pisi isolates in all experiments, the isolate refreshment/multiplication procedures before inoculation may have alter the fungi genetic background of each experiment.

For powdery mildew resistance, two potential candidate resistance genes co-localized with the QTL identified in LG II. One locus corresponds to a “dihydroorotase”. This enzyme is involved in the pirimidine metabolism (Giermann et al., 2002), that, in its turn, is involved in the production of secondary metabolites (Brown and Turan, 1995). Secondary metabolites are often involved in defence, since many possess antimicrobial properties (Stintzi et al., 1993). The other potential candidate gene is the “protein trichome birefringence”, described as being involved in cellulose biosynthesis in Arabidopsis thaliana (Bischoff et al., 2010). Alterations in cellulose synthesis can be monitored by the cell, activating a cascade of signalling events

208 Disease resistance QTL mapping and Synteny studies in L. cicera ______that triggers lignification and defence responses (Caño-Delgado et al., 2003).

Within the confidence intervals of the two QTLs for rust resistance only one molecular marker was functionally annotated, as “MAG2-interacting protein 3”. This protein is involved in protein trafficking between the endoplasmic reticulum (ER) and Golgi in Arabidopsis (Li et al., 2013). The role of ER in defence through protein trafficking and the formation of ER bodies in plants was already postulated (Hayashi et al., 2001; Yamada et al., 2011). ER bodies are located specifically in epidermal cells and contain stress- inducible proteases. These bodies fuse with each other and with vacuoles during stress conditions. Proteases when in vacuoles become active and are discharged to the intracellular space in order to inhibit pathogen development and induce cell death in adjacent cells (Hatsugai et al., 2009).

In our study, the QTL for powdery mildew (qpmres1) resistance is located in a M. truncatula’s syntenic region containing a homologous sequence to the er1 gene (central region of the chromosome 4). The gene er1, identified as PsMLO1 (Humphry et al., 2011), may confer complete or incomplete resistance to powdery mildew in pea depending on the environment (reviewed by Rubiales et al. (2009)). Interestingly this gene was detected up-regulated in the susceptible genotype and down regulated in the partially resistant genotype in our transcriptomics dataset in response to rust (Chapter 5). Therefore, the role of PsMLO1 should be further analysed in the future, also for rust resistance response in legumes.

This study has provided a number of significant genetic outcomes for L. cicera and legume genomics in general. More specifically, the developed linkage map can be used as a tool to

209 Chapter 6 ______localize gene governing other desirable traits. In addition, the already detected resistance QTLs will support the selection of interesting regions for future fine mapping, increasing the potential of precision resistance breeding through Marker Assisted Selection (MAS). As a final future goal, pyramiding of different resistance gene combinations using MAS will result in a more efficient development of new more durable resistant cultivars.

6.6. Acknowledgments

NFA carried out the phenotyping, sample processing for genotyping, performed the data analysis, developed the EST-SSRs primers and the SNP assay, and drafted the manuscript. MLA participated in the construction of the linkage map, QTL mapping and statistical analysis. STL participated in the results generation on the EST-SSR genotyping and data analysis. DR developed de RIL population and revised the manuscript critically. MCVP coordinated the study and participated to the drafting and revision of the manuscript. NFA would like to thank the CRF-INIA, Madrid, Spain, for supplying the genotypes and Professor Zlatko Šatović for the relevant suggestions.

6.7. References Agrios, G.N. (2005). Plant Pathology. New York, U.S.A.: Elsevier Academic Press. Almeida, N.F., Leitão, S.T., Caminero, C., Torres, A.M., Rubiales, D., and Vaz Patto, M.C. (2014a). Transferability of molecular markers from major legumes to Lathyrus spp. for their application in mapping and diversity studies. Molecular Biology Reports 41, 269-283. Almeida, N.F., Leitão, S.T., Krezdorn, N., Rotter, B., Winter, P., Rubiales, D., and Vaz Patto, M.C. (2014b). Allelic diversity in

210 Disease resistance QTL mapping and Synteny studies in L. cicera ______

the transcriptomes of contrasting rust-infected genotypes of Lathyrus sativus, a lasting resource for smart breeding. BMC Plant Biology 14, 376. Aubert, G., Morin, J., Jacquin, F., Loridon, K., Quillet, M.C., Petit, A., Rameau, C., Lejeune-Henaut, I., Huguet, T., and Burstin, J. (2006). Functional mapping in pea, as an aid to the candidate gene selection and for investigating synteny with the model legume Medicago truncatula. Theoretical and Applied Genetics 112, 1024-1041. Bischoff, V., Nita, S., Neumetzler, L., Schindelasch, D., Urbain, A., Eshed, R., Persson, S., Delmer, D., and Scheible, W.-R. (2010). TRICHOME BIREFRINGENCE and its homolog AT5G01360 encode plant-specific DUF231 proteins required for cellulose biosynthesis in Arabidopsis. Plant Physiology 153, 590-602. Brown, E.G., and Turan, Y. (1995). Pyrimidine metabolism and secondary product formation - biogenesis of albizziine, 4- hydroxyhomoarginine and 2,3-diaminopropanoic Acid. Phytochemistry 40, 763-771. Campbell, C.G. (1997). Promoting the conservation and use of underutilized and neglected crops. Grass pea. Lathyrus sativus L. Rome, Italy: International Plant Genetic Resources Institute. Caño-Delgado, A., Penfield, S., Smith, C., Catley, M., and Bevan, M. (2003). Reduced cellulose synthesis invokes lignification and defense responses in Arabidopsis thaliana. The Plant Journal 34, 351-362. Cheng, R., Kleinhofs, A., and Ukai, Y. (1998). Method for mapping a partial lethal-factor locus on a molecular-marker linkage map of a backcross and doubled-haploid population. Theoretical and Applied Genetics 97, 293-298. Choi, H.-K., Kim, D., Uhm, T., Limpens, E., Lim, H., Mun, J.-H., Kalo, P., Penmetsa, R.V., Seres, A., Kulikova, O., Roe, B.A., Bisseling, T., Kiss, G.B., and Cook, D.R. (2004a). A sequence-based genetic map of Medicago truncatula and comparison of marker colinearity with M. sativa. Genetics 166, 1463-1502. Choi, H.K., Mun, J.H., Kim, D.J., Zhu, H.Y., Baek, J.M., Mudge, J., Roe, B., Ellis, N., Doyle, J., Kiss, G.B., Young, N.D., and Cook, D.R. (2004b). Estimating genome conservation between crop and model legume species. Proceedings of the National Academy of Sciences of the United States of America 101, 15289-15294.

211 Chapter 6 ______

Churchill, G.A., and Doerge, R.W. (1994). Empirical threshold values for quantitative trait mapping. Genetics 138, 963-971. Duke, J. (1981). Handbook of Legumes of World Economic Importance. New York, U.S.A.: Plennun Press. Fernández-Aparicio, M., Flores, F., and Rubiales, D. (2009). Field response of Lathyrus cicera germplasm to crenate broomrape (Orobanche crenata). Field Crops Research 113, 321-327. Fondevilla, S., Carver, T.L.W., Moreno, M.T., and Rubiales, D. (2006). Macroscopic and histological characterisation of genes er1 and er2 for powdery mildew resistance in pea. European Journal of Plant Pathology 115, 309-321. Fondevilla, S., and Rubiales, D. (2012). Powdery mildew control in pea. A review. Agronomy for Sustainable Development 32, 401-409. Giermann, N., Schröder, M., Ritter, T., and Zrenner, R. (2002). Molecular analysis of de novo pyrimidine synthesis in solanaceous species. Plant Molecular Biology 50, 393-403. Hanbury, C.D., Siddique, K.H.M., Galwey, N.W., and Cocks, P.S. (1999). Genotype-environment interaction for seed yield and ODAP concentration of Lathyrus sativus L. and L. cicera L. in Mediterranean-type environments. Euphytica 110, 45-60. Hatsugai, N., Iwasaki, S., Tamura, K., Kondo, M., Fuji, K., Ogasawara, K., Nishimura, M., and Hara-Nishimura, I. (2009). A novel membrane fusion-mediated plant immunity against bacterial pathogens. Genes & Development 23, 2496-2506. Hayashi, Y., Yamada, K., Shimada, T., Matsushima, R., Nishizawa, N.K., Nishimura, M., and Hara-Nishimura, I. (2001). A proteinase-storing body that prepares for cell death or stresses in the epidermal cells of Arabidopsis. Plant and Cell Physiology 42, 894-899. Humphry, M., Reinstadler, A., Ivanov, S., Bisseling, T., and Panstruga, R. (2011). Durable broad-spectrum powdery mildew resistance in pea er1 plants is conferred by natural loss-of-function mutations in PsMLO1. Molecular Plant Pathology 12, 866-878. Jansen, R.C., and Stam, P. (1994). High resolution of quantitative traits into multiple loci via interval mapping. Genetics 136, 1447-1455. Kaló, P., Seres, A., Taylor, S.A., Jakab, J., Kevei, Z., Kereszt, A., Endre, G., Ellis, T.H.N., and Kiss, G.B. (2004). Comparative mapping between Medicago sativa and Pisum sativum. Molecular Genetics and Genomics 272, 235-246. Kosambi, D.D. (1943). The estimation of a map distance from recombination values. Annals of Eugenics 12, 172-175.

212 Disease resistance QTL mapping and Synteny studies in L. cicera ______

Lander, E.S., and Botstein, D. (1989). Mapping mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121, 185-199. Li, L., Shimada, T., Takahashi, H., Koumoto, Y., Shirakawa, M., Takagi, J., Zhao, X., Tu, B., Jin, H., Shen, Z., Han, B., Jia, M., Kondo, M., Nishimura, M., and Hara-Nishimura, I. (2013). MAG2 and three MAG2-INTERACTING PROTEINs form an ER-localized complex to facilitate storage protein transport in Arabidopsis thaliana. The Plant Journal 76, 781-791. Martín-Sanz, A., Palomo, J., Pérez De La Vega, M., and Caminero, C. (2012). Characterization of Pseudomonas syringae pv. syringae isolates associated with bacterial blight in Lathyrus spp. and sources of resistance. European Journal of Plant Pathology 134, 205-216. Mcdonald, B., and Linde, C. (2002). The population genetics of plant pathogens and breeding strategies for durable resistance. Euphytica 124, 163-180. Niks, R., and Rubiales, D. (2002). Potentially durable resistance mechanisms in plants to specialised fungal pathogens. Euphytica 124, 201-216. O’connell, R.J., and Panstruga, R. (2006). Tête à tête inside a plant cell: establishing compatibility between plants and biotrophic fungi and oomycetes. New Phytologist 171, 699-718. Parlevliet, J.E. (1979). "The co-evolution of host-parasite systems," in Parasites as plant taxonomists, Acta Universitatis Upsaliensis XXVI, ed. J. Hedberg.), 39-45. Phan, H.T.T., Ellwood, S.R., Hane, J.K., Ford, R., Materne, M., and Oliver, R.P. (2007). Extensive macrosynteny between Medicago truncatula and Lens culinaris ssp. culinaris. Theoretical and Applied Genetics 114, 549-558. Rubiales, D., Castillejo, M.A., Madrid, E., Barilli, E., and Rispail, N. (2011). Legume breeding for rust resistance: lessons to learn from the model Medicago truncatula. Euphytica 180, 89-98. Rubiales, D., Fernández-Aparicio, M., Moral, A., Barilli, E., Sillero, J.C., and Fondevilla, S. (2009). Disease resistance in pea (Pisum sativum L.) types for autumn sowings in Mediterranean environments. Czech Journal of Genetics and Plant Breeding 45, 135-142. Sillero, J.C., Fondevilla, S., Davidson, J., Patto, M.C.V., Warkentin, T.D., Thomas, J., and Rubiales, D. (2006). Screening techniques and sources of resistance to rusts and mildews in grain legumes. Euphytica 147, 255-272. Stakman, E.C., Stewart, D.M., and Loegering, W.Q. (1962). Identification of physiologic races of Puccinia graminis var.

213 Chapter 6 ______

Tritici. United States Department of Agriculture, Agricultural Research Service. Staples, R.C. (2000). Research on the rust fungi during the twentieth century. Annual Review of Phytopathology 38, 49-69. Stintzi, A., Heitz, T., Prasad, V., Wiedemann-Merdinoglu, S., Kauffmann, S., Geoffroy, P., Legrand, M., and Fritig, B. (1993). Plant ‘pathogenesis-related’ proteins and their role in defense against pathogens. Biochimie 75, 687-706. Stuber, C.W., Edwards, M.D., and Wendel, J.F. (1987). Molecular marker-facilitated investigations of quantitative trait loci in maize. II. Factors influencing yield and its component traits. Crop Science 27, 639-648. Torjek, O., Witucka-Wall, H., Meyer, R.C., Von Korff, M., Kusterer, B., Rautengarten, C., and Altmann, T. (2006). Segregation distortion in Arabidopsis C24/Col-0 and Col-0/C24 recombinant inbred line populations is due to reduced fertility caused by epistatic interaction of two loci. Theoretical and Applied Genetics 113, 1551-1561. Torres, A.M., Weeden, N.F., and Martin, A. (1993). Linkage among isozyme, RFLP and RAPD markers in Vicia faba. Theoretical and Applied Genetics 85, 937-945. Van Ooijen, J.W. (2006). JoinMap ® 4, Software for the calculation of genetic linkage maps in experimental populations. Kyazma B.V., Wageningen, Netherlands. Van Ooijen, J.W., Boer, M.P., Jansen, R.C., and Maliepaard, C. (2002). MapQTL® 4.0, Software for the calculation of QTL positions on genetic maps. Plant Research International, Wageningen, the Netherlands. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2006a). Characterization of resistance to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus sativus. Plant Breeding 125, 308-310. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2007). Resistance reaction to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus cicera from Iberian origin. Genetic Resources and Crop Evolution 54, 1517-1521. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2009). Pre and posthaustorial resistance to rusts in Lathyrus cicera L. Euphytica 165, 27-34. Vaz Patto, M.C., and Rubiales, D. (2009). Identification and characterization of partial resistance to rust in a germplasm collection of Lathyrus sativus L. Plant Breeding 128, 495-500.

214 Disease resistance QTL mapping and Synteny studies in L. cicera ______

Vaz Patto, M.C., and Rubiales, D. (2014). Resistance to rust and powdery mildew in Lathyrus crops. Czech Journal of Genetics and Plant Breeding 50, 116-122. Vaz Patto, M.C., Skiba, B., Pang, E.C.K., Ochatt, S.J., Lambein, F., and Rubiales, D. (2006b). Lathyrus improvement for resistance against biotic and abiotic stresses: From classical breeding to marker assisted selection. Euphytica 147, 133- 147. Voorrips, R.E. (2002). MapChart: Software for the graphical presentation of linkage maps and QTLs. Journal of Heredity 93, 77-78. Weeden, N.F., Muehlbauer, F.J., and Ladizinsky, G. (1992). Extensive conservation of linkage relationships between pea and lentil genetic maps. Journal of Heredity 83, 123-129. Xu, Y., Zhu, L., Xiao, J., Huang, N., and Mccouch, S.R. (1997). Chromosomal regions associated with segregation distortion of molecular markers in F2, backcross, doubled haploid, and recombinant inbred populations in rice (Oryza sativa L.). Molecular and General Genetics 253, 535-545. Yamada, K., Hara-Nishimura, I., and Nishimura, M. (2011). Unique defense strategy by the endoplasmic reticulum body in plants. Plant and Cell Physiology 52, 2039-2049.

215

Chapter 7

General discussion

217

General discussion ______

7.1. Project overview

In recent years Lathyrus spp. agronomical potential has been advocated due to their multipurpose grain and forage usage. These species have a hardy and penetrating root system allowing them to grown on a wide range of soil types, including very poor soils and heavy clays (Campbell, 1997). Also Lathyrus spp. are low production cost grain and forage legumes, adapted to low rainfall environments and are considerable potential high quality and cheap protein sources (Hanbury et al., 2000). As legume species, Lathyrus spp., require less fertilizer inputs, since rhizobial bacteria live symbiotically within their root nodules and fix atmospheric nitrogen in a form that can be used by plants (Brewin, 2004). Within the Lathyrus genus, Lathyrus sativus and L. cicera, the two species focus of this thesis, were described as showing resistance to several legume pathogens, such as rust, powdery mildew, ascochyta blight and crenate broomrape (Rubiales et al., 2015).

Earlier investigations on the characterization of such resistance mechanisms against rust (U. pisi) and powdery mildew (E. pisi) were developed by an international collaboration between Portuguese and Spanish teams (Vaz Patto et al., 2006; 2007; 2009; Vaz Patto and Rubiales, 2009). In these studies, an Iberian L. sativus and L. cicera germplasm collection was evaluated for disease resistance, both macro- and microscopically. In both species the resistance against rust and powdery mildew was due to a high frequency of early abortion of the colonies, a reduction in the number of haustoria per colony and the reduction in the intercellular growth of infection hyphae. These phenotypes are typical examples of pre-haustorial resistance with no associated necrosis (Vaz Patto and Rubiales, 2014).

219 Chapter 7 ______

Studies on Ascochyta lathyri resistance in these species are scarce. Previous studies regarding ascochyta blight in Lathyrus spp. were performed using the pea pathogen Didymella pinodes (Gurung et al., 2002; Skiba et al., 2004b; a; 2005). Lathyrus spp., were shown to be significantly more resistant to ascochyta blight than pea, and therefore, findings in these underused species are valuable also to be transferred to close related, economically important crops like pea.

The work described on the present thesis aims to deepen the understanding of the genetics and defence mechanisms, of two Lathyrus spp., to three of the most important foliar diseases in legumes (Rubiales et al., 2015). With that purpose we analysed contrasting genotypes of both L. sativus and L. cicera in what concerns their infection reaction to rust and powdery mildew, and a resistant L. sativus genotype reaction against ascochyta blight. Due to their genetic closeness to other cool season grain legumes, Lathyrus species may share genes, physiological processes and defence mechanisms. Therefore, this knowledge will be useful to several other legume species. Another outcome of this work was the development of new molecular tools, such as the RNA-Seq libraries important for the transcriptome analysis and the development of molecular markers; the development of the first linkage map for L. cicera and the subsequent use of this map to locate rust and powdery mildew resistance QTLs. These new resources will facilitate future research, allowing performing precision breeding on these species, advancing them into the “omics” era.

220 General discussion ______

7.2. Molecular marker development

The first attempt, in this PhD thesis, to obtain molecular markers for Lathyrus spp. was the cross amplification of microsatellites (SSR), intron-targeted amplified polymorphism (ITAP) markers, defence related (DR) genes and resistance gene analogs (RGA) from related legume species (Chapter 2). These markers were tested in our segregating populations L. sativus and L. cicera parental lines with different objectives:

1) Access the transferability of the different markers.

2) Test their use for diversity studies.

3) Identify markers suitable for mapping the segregant progenies and for legume species synteny studies.

Regarding transferability, pea EST-SSRs were significantly more transferable than pea gSSRs. This is not surprising, due to their intrinsic nature. EST-SSRs are exclusively located in coding regions, the more conserved regions of the genome, and gSSRs are distributed indiscriminately throughout the genome, spanning also non-coding regions.

The transferability of the other markers used (ITAP, DR, RGA) was high, mainly because their primers target exons. In the case of the ITAPs their primer pairs flanks intronic regions. This aspect is important for developing polymorphic markers, since in this way, the primer region is conserved, but the region amplified might be more diverse due to its non-coding nature.

This first strategy revealed to be inefficient for one of this thesis main purpose, which was to obtain a large number of cross-species amplified molecular markers useful for genetic mapping of our L. sativus and L. cicera RIL populations. In total, the amplification of

221 Chapter 7 ______heterologous markers in the L. cicera parental lines yielded only five polymorphic ITAP markers. Despite this, a satisfactory amount of markers were obtained for diversity studies in our Lathyrus spp. Iberian collection. This study in particular clearly separated the two species (L. sativus and L. cicera) and most of the individuals within each species.

New attempts for the development of molecular markers were later performed profiting from the L. sativus and L. cicera transcriptomic studies of the differential expression upon inoculation with rust. Several EST-SSRs and single nucleotide polymorphism (SNP) markers were developed by this approach, specifically for those species (Chapters 3 and 5). Since different cDNA libraries were created for each genotype separately, allowing comparison between them, this marker development was targeted for polymorphic sites, increasing the probability of obtaining polymorphic markers very useful for diversity and mapping studies has demonstrated by the inclusion of 29 from the 57 tested EST-SSR on the first L. cicera linkage map developed on Chapter 6 . As expected, this RNA-Seq based molecular marker development proved to be much more efficient than the cross-species amplification, since about 60% of the predicted EST-SSRs were experimentally confirmed to be polymorphic in the same parental genotypes, and suitable for mapping. However, approximately 13% of the developed markers were monomorphic between the parental genotypes. These markers were not useful for the previous mapping purposes, but may be suitable for diversity studies or linkage mapping in more diverse germplasm.

Also with our RNA-Seq data, we compared homolog transcripts from the parental genotypes to search for SNPs. A set of 768 predicted homozygous SNPs were then used for mapping the L. cicera RILs (Chapter 6). From those, 399 SNPs were suitable to screen the RIL

222 General discussion ______population (amplified and were polymorphic between the parental lines), in order to saturate their linkage map. These SNPs were chosen taking in account their homology with the Medicago truncatula genome. This allowed that after the development of the L. cicera linkage map, it was possible to perform comparative studies between these species (L. cicera and M. truncatula).

7.3. Development of the first Lathyrus cicera linkage map

In order to increase the available molecular tools to L. cicera, we developed the first linkage map for this species, based on a recombinant inbred line (RIL) population (Chapter 6). To achieve this goal, the already described set of new molecular markers proven polymorphic between the two parental lines, were screened on the mapping RIL population. Excluding highly distorted and cosegregating markers, the final map was constituted by 258 molecular markers, distributed along 11 linkage maps, with an average marker density of 2.93 cM/marker.

The development of this molecular tool provides new opportunities for genetic studies in L. cicera, like quantitative trait loci (QTL) mapping or comparative mapping to other species. However, before that there are still improvements that can be performed in this linkage map. Many of the EST-SSRs described in this thesis (Chapter 5), where developed after the construction of the linkage map (Chapter 6), and are not yet included in the map. Their inclusion in this map would increase the marker density and contribute to the ultimate goal to reduce the difference from the obtained number of linkage groups in the linkage map and the actual number of L. cicera chromosomes. In addition, a more dense linkage map will allow a more precise QTL/gene

223 Chapter 7 ______mapping. This linkage map will be also useful for further comparative mapping with other species, using the same approach followed as in the comparative mapping to M. truncatula described in Chapter 6. In that analysis, by comparing the relative position of the L. cicera markers to the homologs in the M. truncatula genome, it was possible to observe that the macrosyntenic organization of the genomes is similar, despite L. cicera’s genome size (6.8 Gbp) being 14x bigger than M. truncatula (465 Mbp). Extensive macrosynteny was also observed in previous studies in the Fabaceae tribe such as when comparing M. truncatula to P. sativum (Aubert et al., 2006) and other closely related legumes comparison such as M. truncatula and M. sativa (Choi et al., 2004), L. culinaris and M. truncatula (Phan et al., 2007) and L. culinaris and P. sativum (Weeden et al., 1992). In other study comparing the synteny between M. sativa and P. sativum (Kaló et al., 2004), one rearrangement event, similar to the observed in the comparison between L. cicera and M. truncatula, was observed. In L. cicera, a linkage group (LG IV) was spanning chromosomes 2 and 6 in M. truncatula. Similarly in Kalo et al. (2004), P. sativum’s LG IV was found homolog to M. sativa’s LG 2 and LG 6. This suggested that chromosomal fission/fusion events involving M. truncatula chromosome 6 an d 2 might be responsible for the reduction of chromosome number between M. truncatula/M. sativa (n=8) and L. cicera/P. sativum (n=7).

7.4. Lathyrus spp. response to Uromyces pisi inoculation

In order to understand the genetic basis of resistance to rust in Lathyrus spp., the RNA-Seq libraries from rust inoculated L. sativus and L. cicera were analysed in Chapter 3 and Chapter 5 respectively,

224 General discussion ______complemented with a rust resistance QTL mapping analysis performed for L. cicera in Chapter 6.

Lathyrus sativus differential response to U. pisi inoculation, between the studied resistant (DS=9%) and partially resistant (DS=30%) genotypes, seems to be related to the expression of MLO and MLO-related genes. Mlo interacts with calmodulin to negatively regulate plant defence and promote susceptibility to powdery mildew (Kim et al., 2002). From the 12 differentially expressed MLO transcripts identified between control and inoculated plants, only one was down- regulated, and only in the resistant L. sativus genotype upon inoculation. Since the loss of function of MLO is related to the thickening of the cell wall at penetration sites (Jørgensen, 1992), this may be a common mechanism in the plant defence against powdery mildew and rust. Additionally, emphasizing the importance of the straightening of the cell wall in rust resistance, three transcripts involved in cellulose biogenesis showed a higher up-regulation in the resistant L. sativus genotype than in the susceptible one. Related to defence response induced by phytohormones, transcripts identified as being involved in the salicylic acid pathway had a contrasting expression profile upon inoculation, with genes up-regulated in the partially resistant L. sativus genotype, and down-regulated in the susceptible L. sativus genotype. These identified genes are good candidates for future gene expression studies in order to validate their roles in L. sativus partial resistance to rust.

The studied L. cicera genotypes depicted a lower level of resistance to rust when comparing to L. sativus, presenting a high susceptible genotype (DS=80%) and a partially resistant genotype (DS=36%). For this species, also a MLO-like transcript was found down-regulated upon inoculation in the partially resistant genotype.

225 Chapter 7 ______

This transcript is homologous to A. thaliana’s AtMLO6 and P. sativum’s PsMLO1, which are known to confer susceptibility to several fungal pathogens in Arabidopsis (Chen et al., 2006) and to powdery mildew in P. sativum (Humphry et al., 2011). Interestingly, this transcript is only present in the RNA-Seq library of the most susceptible L. sativus genotype. Several hypotheses could explain its absence on the partial resistant genotype. The gene could be absent from the transcriptome, caused by a deletion or major mutation event that altered significantly the sequence in the gene region. Other cause could be technical, and despite the enrichment steps is the RNA-Seq library preparation prior to sequencing, the expression levels of this transcript might have been so low that could not be detected.

Other genes that can be involved in the contrasting resistance levels between the studied L. cicera genotypes, are genes involved in plant cell wall degradation signalling response through damage- associated molecular patterns (DAMPs) (Boller and Felix, 2009). QTL mapping for rust resistance (Chapter 6) identified 2 QTLs in linkage group II, together explaining 27% of the phenotypic variation. This supports the polygenic nature of the observed resistance, where only the larger QTLs can be detected, but explaining a small percentage of phenotypic variation. The remaining phenotypic variation is induced by small QTLs that were below the detection thresholds. Under those identified QTL regions only one marker could be functionally annotated. This was a protein involved in the protein trafficking between the endoplasmic reticulum (ER) and Golgi in Arabidopsis, the “MAG2-interacting protein 3” (Li et al., 2013). The role of ER in defence through protein trafficking and the formation of ER bodies specifically in plant epidermal cells was already proposed (Hayashi et al., 2001; Yamada et al., 2011). These bodies contain stress-inducible proteases

226 General discussion ______that fuse with each other and with vacuoles during stress conditions. In vacuoles the proteases become active and are discharged to the intracellular space in order to inhibit pathogen development and induce cell death in surrounding cells (Hatsugai et al., 2009).

7.5. Powdery mildew resistance in Lathyrus cicera

As reviewed by Rubiales et al. (2009), so far identified resistance genes to E. pisi in P. sativum consists of only three genes, two recessive (er1 and er2) and one dominant (Er3). Genes er1 and er2 resistance response depends on the environment, displaying complete or incomplete resistance, while Er3 confers always complete resistance. Histological analyses after inoculation of pea genotypes carrying er1, er2 or Er3 genes, showed that these genes influence different stages of the fungal infection process, from a decrease ability of the fungus to penetrate the host cell wall, to a post-penetration cell death response (or hypersensitive response, HR), which results in the collapse of young fungal colonies (Fondevilla et al., 2005; Fondevilla et al., 2007).

Gene er1 was identified as PsMLO1 (Humphry et al., 2011), that depending on the environment may confer complete or incomplete resistance to powdery mildew (reviewed by Rubiales et al.(2009)). Several er1 homologous sequences are found in the M. truncatula genome, being one of them located in the central region of the chromosome 4. From our synteny study, this region corresponds to the L. cicera linkage group II region where the QTLs for powdery mildew and rust resistance were identified.

Co-located with the identified L. cicera powdery mildew resistance QTL, now by direct map analysis, two candidate genes for

227 Chapter 7 ______resistance were identified (Chapter 6). One of these genes is the “protein trichome birefringence”, described as being involved in cellulose biosynthesis in Arabidopsis thaliana (Bischoff et al., 2010). Alterations in cellulose synthesis can be monitored by the cell, activating a cascade of signalling events that triggers lignification and defence responses (Caño-Delgado et al., 2003). This can be somehow related with the er1 function, since in pea (Fondevilla et al., 2006; Humphry et al., 2011), barley (Buschges et al., 1997), Arabidopsis (Consonni et al., 2006) and tomato (Bai et al., 2008), er1 is associated with the fungus being unable to penetrate the host cells. The other co- located gene is the “dihydroorotase”. This enzyme is involved in the pirimidine metabolism (Giermann et al., 2002), that, in its turn, is involved in the production of secondary metabolites (Brown and Turan, 1995). Many of the secondary metabolites have antimicrobial properties, being involved in the inhibition of hyphae growth, which can explain the restriction of haustoria formation with high percentage of early aborted colonies and reduction of intercellular growth of infection hyphae observed in many L. cicera genotypes upon inoculation with U. pisi (Vaz Patto et al., 2009).

7.6. Lathyrus sativus response to Ascochyta lathyri inoculation

In order to investigate the molecular resistance response of a L. sativus genotype to A. lathyri inoculation, we analysed its transcriptome 24h after inoculation, by deepSuperSAGE. This technique sequence a 26bp tag from each transcript, allowing the quantification of the whole transcriptome using reduced sequencing resources. The RNA-Seq library developed in the scope of this thesis (Chapter 5) was crucial to annotate the generated 26bp tags, since the annotation of those tags using only homologous sequences from other

228 General discussion ______species present in the public databases produced a low number of significant BLAST hits.

One important factor of virulence in a necrotroph, as Ascochyta spp., is the ability of induce cell death. No macroscopic necrotic lesions were observed 15 d.a.i., which was in accordance to the transcriptomic data. The only transcript identified as been involved in ROS production was a peroxidase, but was down-regulated upon inoculation, and several detoxification agents, such as glutathione S-transferase, and a precursor of the carotenoids synthesis pathway, were found up- regulated upon inoculation. Since no HR response was evident, we conclude that the resistance of our L. sativus genotype BGE015746 to ascochyta is quantitative rather than qualitative, as it has been reported in other legume species such as pea (Carrillo et al., 2013), lentil (Tullu et al., 2006), faba bean (Rubiales et al., 2012) and chickpea (Hamwieh et al., 2013) and represents a potentially lasting source of resistance to ascochyta blight (Rubiales et al., 2015).

7.7. Future research and practical applications

Profiting from the molecular tools developed under this PhD thesis, further studies can be planned on L. sativus and L. cicera in order to increase information on those crops, and share knowledge with related legume species. These novel molecular tools can be used for diversity studies, to study the genetic basis of other interesting (complex or not) agronomic traits and supporting future precision breeding activities.

For diversity studies, the molecular markers can be used to clarify the relation among genotypes from different domestication centres, useful information to increase the genetic variability in

229 Chapter 7 ______

Lathyrus spp. breeding. The study of a diverse germplasm collection will also enable association mapping studies. The already developed L. cicera linkage map can be further improved by including the remaining EST-SSRs developed, saturating the map in order to link linkage groups belonging to the same chromosome and increase the precision of QTL/gene location. With this linkage map and respective RIL population, it will be also possible to study the genetic basis of other important agronomical traits. For that the parental genotypes could be evaluated, to search for contrasting phenotypes for interesting agronomic traits, such as biotic/abiotic stress responses, ODAP content, yield and plant architecture. Also other linkage maps for Lathyrus spp. could be developed with these markers, using other segregating populations (RILs or F2:3), such as L. sativus RILs segregating for ODAP content, that is currently being developed by our colleagues at IAS-CSIC, Spain.

The obtained linkage map can be used to perform comparative mapping with other legume and non-legume species in order to share knowledge on similar molecular mechanisms conserved among species. This would be important especially for P. sativum which is the closest legume species to Lathyrus spp. and whose genome is predicted to be available in 2016.

The vast array of novel, resistance-related genomic information presented in this thesis provides a highly valuable resource for future smart breeding approaches in these previously under-researched, valuable legume crops.

230 General discussion ______

7.8. References Aubert, G., Morin, J., Jacquin, F., Loridon, K., Quillet, M.C., Petit, A., Rameau, C., Lejeune-Henaut, I., Huguet, T., and Burstin, J. (2006). Functional mapping in pea, as an aid to the candidate gene selection and for investigating synteny with the model legume Medicago truncatula. Theoretical and Applied Genetics 112, 1024-1041. Bai, Y., Pavan, S., Zheng, Z., Zappel, N.F., Reinstadler, A., Lotti, C., De Giovanni, C., Ricciardi, L., Lindhout, P., Visser, R., Theres, K., and Panstruga, R. (2008). Naturally occurring broad- spectrum powdery mildew resistance in a Central American tomato accession is caused by loss of mlo function. Molecular Plant-Microbe Interactions 21, 30-39. Bischoff, V., Nita, S., Neumetzler, L., Schindelasch, D., Urbain, A., Eshed, R., Persson, S., Delmer, D., and Scheible, W.-R. (2010). TRICHOME BIREFRINGENCE and its homolog AT5G01360 encode plant-specific DUF231 proteins required for cellulose biosynthesis in Arabidopsis. Plant Physiology 153, 590-602. Boller, T., and Felix, G. (2009). A Renaissance of elicitors: Perception of microbe-associated molecular patterns and danger signals by pattern-recognition receptors. Annual Review of Plant Biology 60, 379-406. Brewin, N.J. (2004). Plant cell wall remodelling in the rhizobium– legume symbiosis. Critical Reviews in Plant Sciences 23, 293- 316. Brown, E.G., and Turan, Y. (1995). Pyrimidine metabolism and secondary product formation - biogenesis of albizziine, 4- hydroxyhomoarginine and 2,3-diaminopropanoic Acid. Phytochemistry 40, 763-771. Buschges, R., Hollricher, K., Panstruga, R., Simons, G., Wolter, M., Frijters, A., Van Daelen, R., Van Der Lee, T., Diergaarde, P., Groenendijk, J., Topsch, S., Vos, P., Salamini, F., and Schulze- Lefert, P. (1997). The barley Mlo gene: a novel control element of plant pathogen resistance. Cell 88, 695-705. Campbell, C.G. (1997). Promoting the conservation and use of underutilized and neglected crops. Grass pea. Lathyrus sativus L. Rome, Italy: International Plant Genetic Resources Institute. Caño-Delgado, A., Penfield, S., Smith, C., Catley, M., and Bevan, M. (2003). Reduced cellulose synthesis invokes lignification and defense responses in Arabidopsis thaliana. The Plant Journal 34, 351-362. Carrillo, E., Rubiales, D., Perez-De-Luque, A., and Fondevilla, S. (2013). Characterization of mechanisms of resistance against

231 Chapter 7 ______

Didymella pinodes in Pisum spp. European Journal of Plant Pathology 135, 761-769. Chen, Z., Hartmann, H.A., Wu, M.-J., Friedman, E., Chen, J.-G., Pulley, M., Schulze-Lefert, P., Panstruga, R., and Jones, A. (2006). Expression analysis of the AtMLO gene family encoding plant- specific seven-transmembrane domain proteins. Plant Molecular Biology 60, 583-597. Choi, H.-K., Kim, D., Uhm, T., Limpens, E., Lim, H., Mun, J.-H., Kalo, P., Penmetsa, R.V., Seres, A., Kulikova, O., Roe, B.A., Bisseling, T., Kiss, G.B., and Cook, D.R. (2004). A sequence- based genetic map of Medicago truncatula and comparison of marker colinearity with M. sativa. Genetics 166, 1463-1502. Consonni, C., Humphry, M.E., Hartmann, H.A., Livaja, M., Durner, J., Westphal, L., Vogel, J., Lipka, V., Kemmerling, B., Schulze- Lefert, P., Somerville, S.C., and Panstruga, R. (2006). Conserved requirement for a plant host cell protein in powdery mildew pathogenesis. Nature Genetics 38, 716-720. Fondevilla, S., Avila, C.M., Cubero, J.I., and Rubiales, D. (2005). Response to Mycosphaerella pinodes in a germplasm collection of Pisum spp. Plant Breeding 124, 313-315. Fondevilla, S., Carver, T.L.W., Moreno, M.T., and Rubiales, D. (2006). Macroscopic and histological characterisation of genes er1 and er2 for powdery mildew resistance in pea. European Journal of Plant Pathology 115, 309-321. Fondevilla, S., Torres, A.M., Iacute, Moreno, M., A, T., and Rubiales, D. (2007). Identification of a new gene for resistance to powdery mildew in Pisum fulvum, a wild relative of pea. Breeding Science 57, 181-184. Giermann, N., Schröder, M., Ritter, T., and Zrenner, R. (2002). Molecular analysis of de novo pyrimidine synthesis in solanaceous species. Plant Molecular Biology 50, 393-403. Gurung, A.M., Pang, E.C.K., and Taylor, P.W.J. (2002). Examination of Pisum and Lathyrus species as sources of ascochyta blight resistance for field pea (Pisum sativum). Australasian Plant Pathology 31, 41-45. Hamwieh, A., Imtiaz, M., Hobson, K., and Ahmed Kemal, S. (2013). Genetic diversity of microsatellite alleles located at quantitative resistance loci for ascochyta blight resistance in a global collection of chickpea germplasm. Phytopathologia Mediterranea 52, 183-191. Hanbury, C.D., White, C.L., Mullan, B.P., and Siddique, K.H.M. (2000). A review of the potential of Lathyrus sativus L. and L. cicera L. grain for use as animal feed. Animal Feed Science and Technology 87, 1-27.

232 General discussion ______

Hatsugai, N., Iwasaki, S., Tamura, K., Kondo, M., Fuji, K., Ogasawara, K., Nishimura, M., and Hara-Nishimura, I. (2009). A novel membrane fusion-mediated plant immunity against bacterial pathogens. Genes & Development 23, 2496-2506. Hayashi, Y., Yamada, K., Shimada, T., Matsushima, R., Nishizawa, N.K., Nishimura, M., and Hara-Nishimura, I. (2001). A proteinase-storing body that prepares for cell death or stresses in the epidermal cells of Arabidopsis. Plant and Cell Physiology 42, 894-899. Humphry, M., Reinstadler, A., Ivanov, S., Bisseling, T., and Panstruga, R. (2011). Durable broad-spectrum powdery mildew resistance in pea er1 plants is conferred by natural loss-of-function mutations in PsMLO1. Molecular Plant Pathology 12, 866-878. Jørgensen, I.H. (1992). Discovery, characterization and exploitation of Mlo powdery mildew resistance in barley. Euphytica 63, 141- 152. Kaló, P., Seres, A., Taylor, S.A., Jakab, J., Kevei, Z., Kereszt, A., Endre, G., Ellis, T.H.N., and Kiss, G.B. (2004). Comparative mapping between Medicago sativa and Pisum sativum. Molecular Genetics and Genomics 272, 235-246. Kim, M.C., Panstruga, R., Elliott, C., Muller, J., Devoto, A., Yoon, H.W., Park, H.C., Cho, M.J., and Schulze-Lefert, P. (2002). Calmodulin interacts with MLO protein to regulate defence against mildew in barley. Nature 416, 447-451. Li, L., Shimada, T., Takahashi, H., Koumoto, Y., Shirakawa, M., Takagi, J., Zhao, X., Tu, B., Jin, H., Shen, Z., Han, B., Jia, M., Kondo, M., Nishimura, M., and Hara-Nishimura, I. (2013). MAG2 and three MAG2-INTERACTING PROTEINs form an ER-localized complex to facilitate storage protein transport in Arabidopsis thaliana. The Plant Journal 76, 781-791. Phan, H.T.T., Ellwood, S.R., Hane, J.K., Ford, R., Materne, M., and Oliver, R.P. (2007). Extensive macrosynteny between Medicago truncatula and Lens culinaris ssp. culinaris. Theoretical and Applied Genetics 114, 549-558. Rubiales, D., Avila, C.M., Sillero, J.C., Hybl, M., Narits, L., Sass, O., and Flores, F. (2012). Identification and multi-environment validation of resistance to Ascochyta fabae in faba bean (Vicia faba). Field Crops Research 126, 165-170. Rubiales, D., Fernández-Aparicio, M., Moral, A., Barilli, E., Sillero, J.C., and Fondevilla, S. (2009). Disease resistance in pea (Pisum sativum L.) types for autumn sowings in Mediterranean environments. Czech Journal of Genetics and Plant Breeding 45, 135-142.

233 Chapter 7 ______

Rubiales, D., Fondevilla, S., Chen, W., Gentzbittel, L., Higgins, T.J.V., Castillejo, M.A., Singh, K.B., and Rispail, N. (2015). Achievements and challenges in legume breeding for pest and disease resistance. Critical Reviews in Plant Sciences 34, 195- 236. Skiba, B., Ford, R., and Pang, E.C.K. (2004a). Construction of a linkage map based on a Lathyrus sativus backcross population and preliminary investigation of QTLs associated with resistance to ascochyta blight. Theoretical and Applied Genetics 109, 1726-1735. Skiba, B., Ford, R., and Pang, E.C.K. (2004b). Genetics of resistance to Mycosphaerella pinodes in Lathyrus sativus. Australian Journal of Agricultural Research 55, 953–960. Skiba, B., Ford, R., and Pang, E.C.K. (2005). Construction of a cDNA library of Lathyrus sativus inoculated with Mycosphaerella pinodes and the expression of potential defence-related expressed sequence tags (ESTs). Physiological and Molecular Plant Pathology 66, 55-67. Tullu, A., Tar'an, B., Breitkreutz, C., Banniza, S., Warkentin, T.D., Vandenberg, A., and Buchwaldt, L. (2006). A quantitative-trait locus for resistance to ascochyta blight [Ascochyta lentis] maps close to a gene for resistance to anthracnose [Colletotrichum truncatum] in lentil. Canadian Journal of Plant Pathology 28, 588-595. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2006). Characterization of resistance to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus sativus. Plant Breeding 125, 308-310. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2007). Resistance reaction to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus cicera from Iberian origin. Genetic Resources and Crop Evolution 54, 1517-1521. Vaz Patto, M.C., Fernández-Aparicio, M., Moral, A., and Rubiales, D. (2009). Pre and posthaustorial resistance to rusts in Lathyrus cicera L. Euphytica 165, 27-34. Vaz Patto, M.C., and Rubiales, D. (2009). Identification and characterization of partial resistance to rust in a germplasm collection of Lathyrus sativus L. Plant Breeding 128, 495-500. Vaz Patto, M.C., and Rubiales, D. (2014). Resistance to rust and powdery mildew in Lathyrus crops. Czech Journal of Genetics and Plant Breeding 50, 116-122. Weeden, N.F., Muehlbauer, F.J., and Ladizinsky, G. (1992). Extensive conservation of linkage relationships between pea and lentil genetic maps. Journal of Heredity 83, 123-129.

234 General discussion ______

Yamada, K., Hara-Nishimura, I., and Nishimura, M. (2011). Unique defense strategy by the endoplasmic reticulum body in plants. Plant and Cell Physiology 52, 2039-2049.

235